Secure user worker operations

Pass workers/results repos and quorum in the postgres integration
test, matching the constructor change that added quorum voting.

.github/workflows/python.yml

@@ -0,0 +1,28 @@
+name: python
+
+on:
+  push:
+    paths:
+      - "scimesh/**"
+      - "tests/**"
+      - "pyproject.toml"
+      - ".github/workflows/python.yml"
+  pull_request:
+    paths:
+      - "scimesh/**"
+      - "tests/**"
+      - "pyproject.toml"
+      - ".github/workflows/python.yml"
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v4
+      - uses: actions/setup-python@v5
+        with:
+          python-version: "3.11"
+          cache: pip
+      - run: python -m pip install --upgrade pip
+      - run: python -m pip install -e '.[dev]'
+      - run: pytest -q

.github/workflows/users.yml

@@ -0,0 +1,66 @@
+name: users
+
+on:
+  push:
+    paths:
+      - "users/**"
+      - ".github/workflows/users.yml"
+  pull_request:
+    paths:
+      - "users/**"
+      - ".github/workflows/users.yml"
+
+defaults:
+  run:
+    working-directory: users
+
+jobs:
+  test:
+    runs-on: ubuntu-latest
+
+    services:
+      postgres:
+        image: postgres:16-alpine
+        env:
+          POSTGRES_USER: scimesh
+          POSTGRES_PASSWORD: scimesh
+          POSTGRES_DB: scimesh_users
+        ports:
+          - 5432:5432
+        options: >-
+          --health-cmd "pg_isready -U scimesh"
+          --health-interval 5s
+          --health-timeout 3s
+          --health-retries 10
+
+    env:
+      TEST_DATABASE_URL: postgres://scimesh:scimesh@localhost:5432/scimesh_users?sslmode=disable
+
+    steps:
+      - uses: actions/checkout@v4
+
+      - uses: actions/setup-go@v5
+        with:
+          go-version-file: users/go.mod
+          cache-dependency-path: users/go.sum
+
+      - name: go vet
+        run: go vet ./...
+
+      - name: gofmt
+        run: test -z "$(gofmt -l .)" || (gofmt -l . && exit 1)
+
+      - name: unit tests (race)
+        run: go test -race ./...
+
+      - name: lint
+        run: go run github.com/golangci/golangci-lint/v2/cmd/golangci-lint@v2.12.2 run --build-tags=integration ./...
+
+      - name: install migrate CLI
+        run: go install -tags 'postgres' github.com/golang-migrate/migrate/v4/cmd/migrate@v4.17.1
+
+      - name: apply migrations
+        run: migrate -path migrations -database "$TEST_DATABASE_URL" up
+
+      - name: integration tests
+        run: go test -tags=integration ./internal/storage/postgres/ -v

.gitignore

@@ -11,3 +11,8 @@ results/
 *_similarities.csv
 test_results.csv
 test_structures/
+
+# Local coordinator-worker execution state
+worker-data*/
+scimesh-worker-data/
+coordinator/.demo/

Makefile

@@ -0,0 +1,24 @@
+.DEFAULT_GOAL := help
+
+.PHONY: help demo-ui demo-down demo-logs
+
+help:
+	@printf '%s\n' \
+	  'SciMesh developer commands:' \
+	  '  make demo-ui                 Start the local UI pipeline demo with 2 workers.' \
+	  '  make demo-ui WORKERS=3       Start the demo with 3 local workers.' \
+	  '  make demo-logs               Follow coordinator logs for the demo.' \
+	  '  make demo-down               Stop demo containers and workers.' \
+	  '' \
+	  'After make demo-ui: open http://localhost:18080/ui (operator / demo-ui-secret).'
+
+# Convenient entry points from the repository root. Extra settings are passed
+# through, for example: make demo-ui WORKERS=3
+demo-ui:
+	$(MAKE) -C coordinator demo-ui
+
+demo-down:
+	$(MAKE) -C coordinator demo-down
+
+demo-logs:
+	$(MAKE) -C coordinator demo-logs

PLAN.md

@@ -66,7 +66,8 @@ Coordinator reducer -> final artifact -> download/status API
 
 - cloud object storage, Kubernetes, autoscaling, and multi-region operation;
 - arbitrary shell commands sent by coordinator to workers;
-- user accounts, multi-tenancy, billing, or sophisticated authorization;
+- user accounts, multi-tenancy, billing, or sophisticated authorization
+  (planned after the first release in CTX-15);
 - GPU scheduling and multiprocessing inside a worker;
 - Docker as a required runtime dependency;
 - video/CV processing implementation;
@@ -820,6 +821,71 @@ CTX-09 enables final result downloads.
 - failure/retry scenarios have automated coverage;
 - README contains architecture diagram, security caveat, and troubleshooting.
 
+### CTX-13 — In-worker CPU parallelism
+
+**Goal:** Allow a worker to use a bounded, configured number of CPU threads or
+processes while preserving the existing one-task-per-lease coordinator model.
+
+**Depends on:** CTX-12.
+
+**Acceptance criteria:**
+
+- worker concurrency is an explicit configuration value with a safe default of
+  one;
+- a task's internal parallel execution has bounded memory and does not build a
+  dense N×N similarity matrix;
+- CPU-parallel `similarity-search` and `similarity-graph` outputs match the
+  single-threaded local reference byte-for-byte where ordering is observable;
+- result ordering is deterministic across worker counts and block sizes;
+- cancellation, lease loss, and worker failure stop child work safely and do
+  not report a successful result;
+- benchmarks and tests cover one-worker and multi-worker configurations.
+
+### CTX-14 — GPU-accelerated workload execution
+
+**Goal:** Add an optional GPU execution backend for supported molecular
+workloads, while retaining the validated CPU implementation as the reference
+and fallback.
+
+**Depends on:** CTX-13.
+
+**Acceptance criteria:**
+
+- GPU capability and backend version are advertised explicitly by a worker;
+- the coordinator schedules GPU work only to compatible workers and CPU-only
+  workers continue to claim CPU tasks;
+- unsupported hardware, unavailable drivers, and GPU execution errors produce
+  sanitized failures or a documented CPU fallback;
+- GPU results match the CPU reference within a documented, tested numerical
+  tolerance and preserve deterministic output ordering;
+- GPU memory use is bounded and no dense N×N similarity matrix is created;
+- CPU-only CI verifies backend selection and contract behavior, with GPU
+  integration tests documented for compatible runners.
+
+### CTX-15 — User Service and access control
+
+**Goal:** Introduce a dedicated User Service for user identity and access
+control, without coupling workers to user credentials or moving scientific
+workload logic into the service.
+
+**Depends on:** CTX-12.
+
+**Acceptance criteria:**
+
+- the service has a versioned, documented API in
+  [`docs/user-service-api-contract.md`](docs/user-service-api-contract.md) and
+  owns user identity data;
+- credentials and authentication tokens are stored and handled securely; they
+  are never logged or exposed to workers;
+- authenticated identity is propagated to coordinator requests through an
+  explicit, validated boundary;
+- authorization restricts access to jobs and artifacts to the intended user or
+  project;
+- unauthenticated, expired-token, and cross-user access attempts have
+  automated failure tests;
+- the existing single-operator demo remains usable through a documented local
+  development configuration.
+
 ---
 
 ## 10. Suggested assignment bundles
@@ -948,12 +1014,15 @@ Before merging a task, reviewer checks:
 Do not start these before CTX-12 is accepted.
 
 - Replace local artifact storage with S3/MinIO behind an `ArtifactStore` API.
-- Add worker labels/capacity-aware scheduling and concurrency > 1.
+- Add worker labels and capacity-aware scheduling.
+- Implement CTX-13 for bounded in-worker CPU parallelism.
+- Implement CTX-14 for optional GPU-accelerated workload execution.
+- Implement CTX-15 for the User Service and authenticated user/project access.
 - Add cancellation propagation to workers.
 - Add image outputs and final PDF reporting to job artifacts.
 - Add CV/video workloads using the same planner/runner/reducer contract.
 - Add observability export (Prometheus/OpenTelemetry).
-- Add per-user/project authorization and signed artifact URLs.
+- Add signed artifact URLs.
 - Add shard caching and content-addressed input deduplication.
 - Add job priority and fair scheduling.
 - Add a CLI for submitting and monitoring remote jobs.

README.md

@@ -1,10 +1,11 @@
 # SciMesh
 
 SciMesh is a scientific-workload framework for molecular datasets. Its public CLI
-currently runs exact similarity search and sparse similarity-graph construction
-locally in one Python process; it creates no dense similarity matrix. A Python
-Worker client and the planned Go/PostgreSQL coordinator contract are tracked in
-the repository, but distributed execution is not available yet; see
+runs exact similarity search and sparse similarity-graph construction locally in
+one Python process; it creates no dense similarity matrix. The Go/PostgreSQL
+coordinator and Python worker can run a shard-based `similarity-search`
+pipeline locally. After every shard succeeds, the coordinator deterministically
+merges its candidates into one final global top-k CSV. See
 [`STATUS.md`](STATUS.md).
 
 The ChEMBL TSV database is intentionally not included in this repository. Download it separately and pass its path to the commands below. The expected columns are `chembl_id` and `canonical_smiles`.
@@ -43,6 +44,27 @@ scimesh similarity-search --help
 scimesh similarity-graph --help
 ```
 
+## Manual pipeline demo
+
+To inspect the coordinator, Web UI, and distributed `similarity-search`
+pipeline by hand, install development dependencies once and start the isolated
+demo from the repository root:
+
+```bash
+python3 -m venv .venv
+.venv/bin/pip install -e '.[dev]'
+make demo-ui
+```
+
+Open `http://localhost:18080/ui` and sign in with username `operator` and
+password `demo-ui-secret`. The command starts PostgreSQL, the coordinator, and
+two local reference workers. Upload a small ChEMBL TSV, then use the job page
+to follow shard progress, inspect bounded **Preview CSV** results, and see a
+live processing-speed chart in shards per minute. To change the worker count,
+run `make demo-ui WORKERS=3`; stop everything with `make demo-down`.
+
+Run `make help` to display these commands in the terminal.
+
 ## Similarity search
 
 `similarity-search` finds the top-k molecules most similar to a query. The query is supplied either by ChEMBL ID or by SMILES. It uses Morgan fingerprints with `radius=2` and `fpSize=2048`, Tanimoto similarity, streaming TSV reads, and a bounded heap. Invalid SMILES and the query molecule are skipped.
@@ -110,3 +132,11 @@ pytest
 ```
 
 The package separates common dataset parsing and fingerprints from independent workloads. Add future workloads through the workload registry without changing the main CLI.
+
+## Team
+
+- [Emil](https://github.com/emil28092005) — Project Lead
+- [Kristina](https://github.com/kristtma) — Tech Lead
+- [Veniamin](https://t.me/Veniamin_Kt) — Scientific Lead
+- [Arkhip](https://github.com/hIpa-ussr) — Programmer
+- [Reranchik](https://github.com/RERAN4K) — Programmer

STATUS.md

@@ -1,7 +1,7 @@
 # SciMesh Status
 
-**Updated:** 2026-07-23  
-**Branch baseline:** `main` at `b4a89dd` (coordinator merge)
+**Updated:** 2026-07-27
+**Branch baseline:** `main` at `f5ead0a` (team and scaling-roadmap documentation)
 
 ## Current state
 
@@ -18,8 +18,13 @@ the reference behaviour for future distributed execution:
 The Go coordinator and its PostgreSQL-backed task lifecycle are implemented:
 registration, atomic claiming, lease renewal, artifact storage, dataset
 chunking, result/failure reporting, and job progress. The Python worker now
-uses the live coordinator contract; its HTTP path was exercised against a real
-Docker PostgreSQL stack on 2026-07-23.
+uses the live coordinator contract. Completed similarity-search shard results
+are reduced once into a checksum-protected final CSV, which is downloadable
+through the coordinator. The full Go checks (including a fresh migration and
+real PostgreSQL smoke test) passed on 2026-07-24.
+
+A User Service is being developed on a separate programmer branch. It is not
+yet merged, reviewed, or integrated with the coordinator/worker contract.
 
 ## Milestone tracker
 
@@ -32,27 +37,29 @@ Docker PostgreSQL stack on 2026-07-23.
 | CTX-04 Worker registry and HTTP API | Implemented | Registration, claim, heartbeat, result, failure, and status endpoints. |
 | CTX-05 Artifact storage | Implemented | Coordinator-owned inputs/results, checksum verification, and upload flow. |
 | CTX-06 Python Worker live-contract alignment | Implemented | Worker completed a real uploaded shard via HTTP on 2026-07-23. |
-| CTX-07 Distributed workload protocol | Not started | Depends on artifact and Worker contracts. |
-| CTX-08 Distributed similarity-search | Not started | Local reference exists. |
-| CTX-09 Reducer and final-result API | Not started | Depends on CTX-07 and CTX-08. |
+| CTX-07 Distributed workload protocol | Implemented | Versioned Python contract models, registry, strict plan validation, and deterministic reduction ordering are in `scimesh/distributed/`. |
+| CTX-08 Distributed similarity-search | Implemented | Python planner resolves `query_id` once, creates deterministic shard plans, worker adapter emits exact partial top-k CSVs/metrics, and reducer matches the local reference. |
+| CTX-09 Reducer and final-result API | Implemented | Atomic `reducing` claim, deterministic coordinator-side top-k reducer, sanitized reducer failure, final artifact persistence, `result_uri`, and final CSV download. |
 | CTX-10 Distributed similarity-graph | Not started | Local reference exists. |
-| CTX-11 Dashboard/operator view | In progress | `feat/web-interface` adds a protected local view: job/task/worker status, dataset upload, diagnostic partial-artifact download, and polling. Final-result reduction remains CTX-09. |
+| CTX-11 Dashboard/operator view | Implemented | Protected live control room: recent-run/worker overview, real pipeline-stage visualization, shard attempts and safe failures, validated similarity-search upload, coordinator artifacts, final-result download, and bounded polling. |
 | CTX-12 Reliability, security, CI | In progress | Unit, race, PostgreSQL integration, and smoke checks exist; CI hardening remains. |
+| CTX-15 User Service and access control | In progress (separate branch) | Proposed implementation is under development; API, security review, tests, and integration are pending. |
 
 ## Next recommended assignment
 
-Assign **CTX-07** to the workload role: define distributed job planning and
-reduction boundaries before implementing distributed search or graph execution.
+Assign **CTX-10** to the distributed-science role: implement deterministic
+block-pair planning and reduction for `similarity-graph`.
 
 ## Known constraints
 
-- Planner/reducer semantics are not implemented; the operator UI labels
-  `partial_result` files as diagnostic and cannot present them as final output.
-  Use the local `scimesh` CLI for complete workload results.
 - The worker/coordinator flow currently accepts both underscore API workload
   names and hyphenated CLI names while the contract is consolidated.
-- A real-stack worker test uses a small `query_smiles` shard. Resolving a
-  `query_id` once and sharing it across shards belongs to CTX-07.
+- A real-stack worker test uses a small `query_smiles` shard. The Python
+  planner resolves `query_id` once and shares `query_smiles`; the upload UI
+  currently accepts `query_smiles` only.
+- The coordinator accepts uploaded distributed jobs only for
+  `similarity-search` with `query_smiles`. It rejects `similarity-graph` until
+  CTX-10 supplies cross-shard pair planning.
 
 ## Update rule
 

coordinator/Dockerfile

@@ -5,7 +5,7 @@
 # build fails with "the --mount option requires BuildKit".
 
 # --- build stage ----------------------------------------------------------
-FROM golang:1.24-alpine AS build
+FROM golang:1.25-alpine AS build
 
 WORKDIR /src
 

coordinator/Makefile

@@ -1,4 +1,6 @@
-.PHONY: build run test test-integration vet lint tidy check migrate-up migrate-down up down down-clean logs ps rebuild psql smoke
+.DEFAULT_GOAL := help
+
+.PHONY: help build run test test-integration vet lint tidy check migrate-up migrate-down up down down-clean logs ps rebuild psql smoke demo-ui demo-down demo-logs
 
 # `check` deliberately uses its own Compose project and host ports. This keeps
 # it from connecting to or replacing a developer's local PostgreSQL instance.
@@ -10,6 +12,59 @@ CHECK_TOKEN ?= dev-token
 CHECK_DATABASE_URL ?= postgres://scimesh:scimesh@localhost:$(CHECK_POSTGRES_PORT)/scimesh?sslmode=disable
 CHECK_COMPOSE = POSTGRES_PORT=$(CHECK_POSTGRES_PORT) COORDINATOR_PORT=$(CHECK_COORDINATOR_PORT) docker compose -p $(CHECK_PROJECT)
 
+# --- local manual demo ---------------------------------------------------
+# A separate project and ports mean this demo cannot collide with the normal
+# `make up` stack or a developer's local PostgreSQL on 5432.
+DEMO_PROJECT ?= scimesh-demo
+DEMO_POSTGRES_PORT ?= 55432
+DEMO_COORDINATOR_PORT ?= 18080
+DEMO_UI_TOKEN ?= demo-ui-secret
+DEMO_WORKER_TOKEN ?= demo-worker-token
+DEMO_WORKERS ?= 2
+# Short public knob for `make demo-ui WORKERS=3`; DEMO_WORKERS remains useful
+# for scripts and backwards-compatible documentation.
+WORKERS ?= $(DEMO_WORKERS)
+DEMO_DIR ?= .demo
+
+help:
+	@printf '%s\n' \
+	  'SciMesh coordinator commands:' \
+	  '  make up / make down          Start or stop the normal coordinator stack.' \
+	  '  make demo-ui [WORKERS=3]     Start isolated UI demo services and local workers.' \
+	  '  make demo-logs               Follow coordinator logs for the UI demo.' \
+	  '  make demo-down               Stop the demo services and workers.' \
+	  '  make test / make vet         Run Go verification.' \
+	  '' \
+	  'Demo UI: http://localhost:18080/ui (login page; admin root@scimesh.local / rootpassword).'
+
+demo-ui:
+	@DEMO_PROJECT="$(DEMO_PROJECT)" \
+	DEMO_POSTGRES_PORT="$(DEMO_POSTGRES_PORT)" \
+	DEMO_COORDINATOR_PORT="$(DEMO_COORDINATOR_PORT)" \
+	DEMO_UI_TOKEN="$(DEMO_UI_TOKEN)" \
+	DEMO_WORKER_TOKEN="$(DEMO_WORKER_TOKEN)" \
+	DEMO_WORKERS="$(WORKERS)" \
+	DEMO_DIR="$(DEMO_DIR)" \
+	./scripts/demo-ui.sh start
+
+demo-down:
+	@DEMO_PROJECT="$(DEMO_PROJECT)" \
+	DEMO_POSTGRES_PORT="$(DEMO_POSTGRES_PORT)" \
+	DEMO_COORDINATOR_PORT="$(DEMO_COORDINATOR_PORT)" \
+	DEMO_UI_TOKEN="$(DEMO_UI_TOKEN)" \
+	DEMO_WORKER_TOKEN="$(DEMO_WORKER_TOKEN)" \
+	DEMO_DIR="$(DEMO_DIR)" \
+	./scripts/demo-ui.sh stop
+
+demo-logs:
+	@DEMO_PROJECT="$(DEMO_PROJECT)" \
+	DEMO_POSTGRES_PORT="$(DEMO_POSTGRES_PORT)" \
+	DEMO_COORDINATOR_PORT="$(DEMO_COORDINATOR_PORT)" \
+	DEMO_UI_TOKEN="$(DEMO_UI_TOKEN)" \
+	DEMO_WORKER_TOKEN="$(DEMO_WORKER_TOKEN)" \
+	DEMO_DIR="$(DEMO_DIR)" \
+	./scripts/demo-ui.sh logs
+
 # --- build / run ---------------------------------------------------------
 build:
 	go build ./...

coordinator/README.md

@@ -76,9 +76,45 @@ UI_AUTH_TOKEN='local-ui-secret' make up
 ```
 
 The UI is disabled by default and never accepts the worker bearer token.
-It shows recent jobs, task/worker state, and the per-job partial artifacts.
-Those files are explicitly diagnostic until the CTX-09 reducer creates a final
-result; the UI does not present them as final scientific output.
+The **control room** shows live workers, recent runs, shard state/attempts,
+safe failures, coordinator artifacts, and the final CSV for completed
+similarity-search jobs. The job page follows the real stages: TSV accepted →
+shards execute → workers return CSVs → `reducing` → final deterministic global
+top-k result. It polls only its own coordinator read-model and never controls
+or exposes worker processes.
+
+For a hands-on run, open `/ui`, choose **New similarity search**, select a
+small ChEMBL-style TSV, then leave one or more `scimesh-worker` processes
+running in separate terminals. The detail page updates every two seconds and
+stops polling after a completed, failed, or cancelled job. Use **Preview CSV**
+to inspect a bounded first page of a partial or completed final result before
+downloading it. The UI never exposes source datasets or shard inputs; partial
+CSVs remain available only as diagnostics.
+
+### One-command manual demo
+
+From the repository root, create the Python environment once, then start a
+self-contained UI demo with two local reference workers:
+
+```sh
+python3 -m venv .venv
+.venv/bin/pip install -e '.[dev]'
+make demo-ui
+```
+
+This uses a separate Docker project and ports `18080` (coordinator) and
+`55432` (PostgreSQL), so it does not conflict with the normal stack. Open
+`http://localhost:18080/ui`, use username `operator` and password
+`demo-ui-secret`, upload a small ChEMBL TSV, and observe the workers process
+it. Change the worker count with `make demo-ui WORKERS=3`; stop all demo
+services and workers with `make demo-down`.
+
+The job page shows a live **Processing speed** graph in completed shards per
+minute. It uses the coordinator snapshots observed by the open browser tab, so
+it is a transparent local-session measurement rather than a persisted metric.
+Use **Preview CSV** before downloading a partial diagnostic or completed final
+result. Run `make help` from either the repository root or this directory for
+the full list of demo commands.
 
 `up` starts three services in order: Postgres waits until `pg_isready` passes, a
 one-shot `migrate` container applies the schema and exits, and only then does the

coordinator/cmd/coordinator/main.go

@@ -9,6 +9,7 @@ import (
 	"syscall"
 
 	"github.com/emil28092005/SciMesh/coordinator/internal/infra"
+	"github.com/emil28092005/SciMesh/coordinator/internal/metrics"
 	"github.com/emil28092005/SciMesh/coordinator/internal/storage/blob"
 	"github.com/emil28092005/SciMesh/coordinator/internal/storage/postgres"
 	httptransport "github.com/emil28092005/SciMesh/coordinator/internal/transport/http"
@@ -66,28 +67,32 @@ func run() error {
 		workerRepo     = postgres.NewWorkerRepo(pool)
 		artifactRepo   = postgres.NewArtifactRepo(pool)
 		uiReadRepo     = postgres.NewUIReadRepo(pool)
+		taskResultRepo = postgres.NewTaskResultRepo(pool)
 	)
 
 	useCases := httptransport.UseCases{
 		RegisterWorker:   usecase.NewRegisterWorker(workerRepo, clk),
 		CreateJob:        usecase.NewCreateJob(jobRepo, taskRepo, tx, clk),
-		SubmitDataset:    usecase.NewSubmitDataset(blobStore, artifactRepo, jobRepo, taskRepo, tx, clk),
-		ClaimTask:        usecase.NewClaimTask(taskRepo, clk, cfg.LeaseDuration),
+		SubmitDataset:    usecase.NewSubmitDataset(blobStore, artifactRepo, jobRepo, taskRepo, tx, clk, cfg.DefaultMaxAttempts),
+		ClaimTask:        usecase.NewClaimTask(taskRepo, jobRepo, workerRepo, tx, clk, cfg.LeaseDuration),
 		RenewLease:       usecase.NewRenewLease(taskRepo, workerRepo, tx, clk, cfg.LeaseDuration),
-		CompleteTask:     usecase.NewCompleteTask(taskRepo, jobRepo, artifactRepo, tx, clk),
-		FailTask:         usecase.NewFailTask(taskRepo, jobRepo, tx, clk),
+		CompleteTask:     usecase.NewCompleteTask(taskRepo, jobRepo, artifactRepo, workerRepo, taskResultRepo, tx, clk, cfg.QuorumSize),
+		ReduceJob:        usecase.NewReduceJob(jobRepo, taskRepo, artifactRepo, blobStore, tx, clk),
+		FailTask:         usecase.NewFailTask(taskRepo, jobRepo, workerRepo, tx, clk),
 		GetJobStatus:     usecase.NewGetJobStatus(jobRepo, taskRepo),
 		CancelJob:        usecase.NewCancelJob(jobRepo, taskRepo, tx, clk),
-		UploadArtifact:   usecase.NewUploadArtifact(taskRepo, artifactRepo, blobStore, clk),
+		UploadArtifact:   usecase.NewUploadArtifact(taskRepo, workerRepo, artifactRepo, blobStore, tx, clk),
 		DownloadArtifact: usecase.NewDownloadArtifact(artifactRepo, blobStore),
+		GetJobResult:     usecase.NewGetJobResult(jobRepo, usecase.NewDownloadArtifact(artifactRepo, blobStore)),
 		GetTaskInput:     usecase.NewGetTaskInput(taskRepo, artifactRepo, blobStore),
 		Dashboard:        usecase.NewDashboard(uiReadRepo),
+		PreviewArtifact:  usecase.NewPreviewArtifact(uiReadRepo, blobStore),
 	}
 
 	// Background reapers are tracked so shutdown can wait for them. Without this
 	// the process would exit mid-UPDATE, and the deferred pool.Close() would pull
 	// connections out from under them.
-	expireLeases := usecase.NewExpireLeases(taskRepo, clk)
+	expireLeases := usecase.NewExpireLeases(taskRepo, jobRepo, tx, clk)
 	markOffline := usecase.NewMarkWorkersOffline(workerRepo, clk, cfg.WorkerOfflineAfter)
 
 	var wg sync.WaitGroup
@@ -105,9 +110,18 @@ func run() error {
 		}(r.name, r.fn)
 	}
 
+	// Business metrics: gauges of tasks/jobs/workers by status, sampled from the
+	// database on every Prometheus scrape.
+	statsRepo := postgres.NewStatsRepo(pool)
+	m := metrics.New()
+	m.RegisterBusiness(func(ctx context.Context) (metrics.Stats, error) {
+		tasks, jobs, workers, err := statsRepo.Counts(ctx)
+		return metrics.Stats{Tasks: tasks, Jobs: jobs, Workers: workers}, err
+	})
+
 	// pool.Ping backs /health: readiness means the database answers, not just
 	// that the process is alive.
-	api := httptransport.NewServer(useCases, log, cfg.RequestTimeout, cfg.HeartbeatInterval, cfg.MaxUploadBytes, pool.Ping)
+	api := httptransport.NewServer(useCases, log, cfg.RequestTimeout, cfg.HeartbeatInterval, cfg.MaxUploadBytes, cfg.JWTSecret, cfg.UserserviceURL, m, pool.Ping)
 	err = infra.RunServer(ctx, log, cfg.Addr, api.Handler(cfg.Token, cfg.UIToken))
 
 	// Shutdown order matters, and defers alone cannot express it (they run

coordinator/docker-compose.monitoring.yml

@@ -0,0 +1,30 @@
+# Demo overlay: Prometheus scrapes the coordinator's /metrics, Grafana shows the
+# provisioned SciMesh dashboard. Merged by scripts/demo-ui.sh with a third -f.
+# Both share the coordinator's compose network, so Prometheus reaches it by name.
+
+services:
+  prometheus:
+    image: prom/prometheus:v2.54.1
+    volumes:
+      - ./monitoring/prometheus.yml:/etc/prometheus/prometheus.yml:ro
+    ports:
+      - "${PROMETHEUS_PORT:-19090}:9090"
+    restart: unless-stopped
+
+  grafana:
+    image: grafana/grafana:11.2.0
+    depends_on:
+      - prometheus
+    environment:
+      GF_SECURITY_ADMIN_USER: admin
+      GF_SECURITY_ADMIN_PASSWORD: ${GRAFANA_PASSWORD:-admin}
+      # Anonymous viewing so the demo dashboard opens without a login.
+      GF_AUTH_ANONYMOUS_ENABLED: "true"
+      GF_AUTH_ANONYMOUS_ORG_ROLE: Viewer
+      GF_USERS_DEFAULT_THEME: dark
+    volumes:
+      - ./monitoring/grafana/provisioning:/etc/grafana/provisioning:ro
+      - ./monitoring/grafana/dashboards:/var/lib/grafana/dashboards:ro
+    ports:
+      - "${GRAFANA_PORT:-13000}:3000"
+    restart: unless-stopped

coordinator/docker-compose.users.yml

@@ -0,0 +1,66 @@
+# Demo overlay: adds the userservice (its own Postgres + migrations) alongside
+# the coordinator and wires the two together with a shared JWT secret, so the
+# operator UI authenticates through userservice login/registration.
+#
+# Used only by scripts/demo-ui.sh, merged onto docker-compose.yml with a second
+# -f. Not part of the plain `make up` stack.
+
+services:
+  postgres-users:
+    image: postgres:16-alpine
+    environment:
+      POSTGRES_USER: ${POSTGRES_USER:-scimesh}
+      POSTGRES_PASSWORD: ${POSTGRES_PASSWORD:-scimesh}
+      POSTGRES_DB: scimesh_users
+    healthcheck:
+      test: ["CMD-SHELL", "pg_isready -U ${POSTGRES_USER:-scimesh} -d scimesh_users"]
+      interval: 5s
+      timeout: 3s
+      retries: 10
+      start_period: 5s
+
+  migrate-users:
+    image: migrate/migrate:v4.17.1
+    depends_on:
+      postgres-users:
+        condition: service_healthy
+    volumes:
+      - ../users/migrations:/migrations:ro
+    command:
+      - -path=/migrations
+      - -database=postgres://${POSTGRES_USER:-scimesh}:${POSTGRES_PASSWORD:-scimesh}@postgres-users:5432/scimesh_users?sslmode=disable
+      - up
+    restart: on-failure
+
+  userservice:
+    build:
+      context: ../users
+    depends_on:
+      postgres-users:
+        condition: service_healthy
+      migrate-users:
+        condition: service_completed_successfully
+    environment:
+      USERSERVICE_ADDR: ":8081"
+      DATABASE_URL: postgres://${POSTGRES_USER:-scimesh}:${POSTGRES_PASSWORD:-scimesh}@postgres-users:5432/scimesh_users?sslmode=disable
+      JWT_SECRET: ${JWT_SECRET}
+      # Seeds the first admin the very first time it boots (idempotent after).
+      BOOTSTRAP_ADMIN_EMAIL: ${BOOTSTRAP_ADMIN_EMAIL:-root@scimesh.local}
+      BOOTSTRAP_ADMIN_PASSWORD: ${BOOTSTRAP_ADMIN_PASSWORD}
+      LOG_LEVEL: ${LOG_LEVEL:-info}
+    ports:
+      - "${USERSERVICE_PORT:-18081}:8081"
+    healthcheck:
+      test: ["CMD", "wget", "-qO-", "http://127.0.0.1:8081/health"]
+      interval: 10s
+      timeout: 3s
+      retries: 3
+      start_period: 5s
+    restart: unless-stopped
+
+  # Turn the coordinator UI into session mode: the same shared secret verifies
+  # userservice tokens locally, and USERSERVICE_URL is where login/register proxy.
+  coordinator:
+    environment:
+      JWT_SECRET: ${JWT_SECRET}
+      USERSERVICE_URL: http://userservice:8081

coordinator/go.mod

@@ -1,23 +1,33 @@
 module github.com/emil28092005/SciMesh/coordinator
 
-go 1.22
+go 1.25.0
 
 require (
 	github.com/Masterminds/squirrel v1.5.4
 	github.com/cenkalti/backoff/v4 v4.3.0
+	github.com/golang-jwt/jwt/v5 v5.3.1
 	github.com/google/uuid v1.6.0
 	github.com/jackc/pgx/v5 v5.6.0
 	github.com/joho/godotenv v1.5.1
+	github.com/prometheus/client_golang v1.19.1
 	gopkg.in/natefinch/lumberjack.v2 v2.2.1
 )
 
 require (
+	github.com/beorn7/perks v1.0.1 // indirect
+	github.com/cespare/xxhash/v2 v2.3.0 // indirect
 	github.com/jackc/pgpassfile v1.0.0 // indirect
 	github.com/jackc/pgservicefile v0.0.0-20221227161230-091c0ba34f0a // indirect
 	github.com/jackc/puddle/v2 v2.2.1 // indirect
 	github.com/lann/builder v0.0.0-20180802200727-47ae307949d0 // indirect
 	github.com/lann/ps v0.0.0-20150810152359-62de8c46ede0 // indirect
+	github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822 // indirect
+	github.com/prometheus/client_model v0.6.2 // indirect
+	github.com/prometheus/common v0.55.0 // indirect
+	github.com/prometheus/procfs v0.21.1 // indirect
 	golang.org/x/crypto v0.17.0 // indirect
-	golang.org/x/sync v0.1.0 // indirect
-	golang.org/x/text v0.14.0 // indirect
+	golang.org/x/sync v0.22.0 // indirect
+	golang.org/x/sys v0.47.0 // indirect
+	golang.org/x/text v0.40.0 // indirect
+	google.golang.org/protobuf v1.36.11 // indirect
 )

coordinator/go.sum

@@ -1,10 +1,18 @@
 github.com/Masterminds/squirrel v1.5.4 h1:uUcX/aBc8O7Fg9kaISIUsHXdKuqehiXAMQTYX8afzqM=
 github.com/Masterminds/squirrel v1.5.4/go.mod h1:NNaOrjSoIDfDA40n7sr2tPNZRfjzjA400rg+riTZj10=
+github.com/beorn7/perks v1.0.1 h1:VlbKKnNfV8bJzeqoa4cOKqO6bYr3WgKZxO8Z16+hsOM=
+github.com/beorn7/perks v1.0.1/go.mod h1:G2ZrVWU2WbWT9wwq4/hrbKbnv/1ERSJQ0ibhJ6rlkpw=
 github.com/cenkalti/backoff/v4 v4.3.0 h1:MyRJ/UdXutAwSAT+s3wNd7MfTIcy71VQueUuFK343L8=
 github.com/cenkalti/backoff/v4 v4.3.0/go.mod h1:Y3VNntkOUPxTVeUxJ/G5vcM//AlwfmyYozVcomhLiZE=
+github.com/cespare/xxhash/v2 v2.3.0 h1:UL815xU9SqsFlibzuggzjXhog7bL6oX9BbNZnL2UFvs=
+github.com/cespare/xxhash/v2 v2.3.0/go.mod h1:VGX0DQ3Q6kWi7AoAeZDth3/j3BFtOZR5XLFGgcrjCOs=
 github.com/davecgh/go-spew v1.1.0/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
 github.com/davecgh/go-spew v1.1.1 h1:vj9j/u1bqnvCEfJOwUhtlOARqs3+rkHYY13jYWTU97c=
 github.com/davecgh/go-spew v1.1.1/go.mod h1:J7Y8YcW2NihsgmVo/mv3lAwl/skON4iLHjSsI+c5H38=
+github.com/golang-jwt/jwt/v5 v5.3.1 h1:kYf81DTWFe7t+1VvL7eS+jKFVWaUnK9cB1qbwn63YCY=
+github.com/golang-jwt/jwt/v5 v5.3.1/go.mod h1:fxCRLWMO43lRc8nhHWY6LGqRcf+1gQWArsqaEUEa5bE=
+github.com/google/go-cmp v0.7.0 h1:wk8382ETsv4JYUZwIsn6YpYiWiBsYLSJiTsyBybVuN8=
+github.com/google/go-cmp v0.7.0/go.mod h1:pXiqmnSA92OHEEa9HXL2W4E7lf9JzCmGVUdgjX3N/iU=
 github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0=
 github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
 github.com/jackc/pgpassfile v1.0.0 h1:/6Hmqy13Ss2zCq62VdNG8tM1wchn8zjSGOBJ6icpsIM=
@@ -21,20 +29,34 @@ github.com/lann/builder v0.0.0-20180802200727-47ae307949d0 h1:SOEGU9fKiNWd/HOJuq
 github.com/lann/builder v0.0.0-20180802200727-47ae307949d0/go.mod h1:dXGbAdH5GtBTC4WfIxhKZfyBF/HBFgRZSWwZ9g/He9o=
 github.com/lann/ps v0.0.0-20150810152359-62de8c46ede0 h1:P6pPBnrTSX3DEVR4fDembhRWSsG5rVo6hYhAB/ADZrk=
 github.com/lann/ps v0.0.0-20150810152359-62de8c46ede0/go.mod h1:vmVJ0l/dxyfGW6FmdpVm2joNMFikkuWg0EoCKLGUMNw=
+github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822 h1:C3w9PqII01/Oq1c1nUAm88MOHcQC9l5mIlSMApZMrHA=
+github.com/munnerz/goautoneg v0.0.0-20191010083416-a7dc8b61c822/go.mod h1:+n7T8mK8HuQTcFwEeznm/DIxMOiR9yIdICNftLE1DvQ=
 github.com/pmezard/go-difflib v1.0.0 h1:4DBwDE0NGyQoBHbLQYPwSUPoCMWR5BEzIk/f1lZbAQM=
 github.com/pmezard/go-difflib v1.0.0/go.mod h1:iKH77koFhYxTK1pcRnkKkqfTogsbg7gZNVY4sRDYZ/4=
+github.com/prometheus/client_golang v1.19.1 h1:wZWJDwK+NameRJuPGDhlnFgx8e8HN3XHQeLaYJFJBOE=
+github.com/prometheus/client_golang v1.19.1/go.mod h1:mP78NwGzrVks5S2H6ab8+ZZGJLZUq1hoULYBAYBw1Ho=
+github.com/prometheus/client_model v0.6.2 h1:oBsgwpGs7iVziMvrGhE53c/GrLUsZdHnqNwqPLxwZyk=
+github.com/prometheus/client_model v0.6.2/go.mod h1:y3m2F6Gdpfy6Ut/GBsUqTWZqCUvMVzSfMLjcu6wAwpE=
+github.com/prometheus/common v0.55.0 h1:KEi6DK7lXW/m7Ig5i47x0vRzuBsHuvJdi5ee6Y3G1dc=
+github.com/prometheus/common v0.55.0/go.mod h1:2SECS4xJG1kd8XF9IcM1gMX6510RAEL65zxzNImwdc8=
+github.com/prometheus/procfs v0.21.1 h1:GljZCt+zSTS+NZq88cyQ1LjZ+RCHp3uVuabBWA5+OJI=
+github.com/prometheus/procfs v0.21.1/go.mod h1:aB55Cww9pdSJVHk0hUf0inxWyyjPogFIjmHKYgMKmtY=
 github.com/stretchr/objx v0.1.0/go.mod h1:HFkY916IF+rwdDfMAkV7OtwuqBVzrE8GR6GFx+wExME=
 github.com/stretchr/testify v1.2.2/go.mod h1:a8OnRcib4nhh0OaRAV+Yts87kKdq0PP7pXfy6kDkUVs=
 github.com/stretchr/testify v1.3.0/go.mod h1:M5WIy9Dh21IEIfnGCwXGc5bZfKNJtfHm1UVUgZn+9EI=
 github.com/stretchr/testify v1.7.0/go.mod h1:6Fq8oRcR53rry900zMqJjRRixrwX3KX962/h/Wwjteg=
-github.com/stretchr/testify v1.8.1 h1:w7B6lhMri9wdJUVmEZPGGhZzrYTPvgJArz7wNPgYKsk=
-github.com/stretchr/testify v1.8.1/go.mod h1:w2LPCIKwWwSfY2zedu0+kehJoqGctiVI29o6fzry7u4=
+github.com/stretchr/testify v1.9.0 h1:HtqpIVDClZ4nwg75+f6Lvsy/wHu+3BoSGCbBAcpTsTg=
+github.com/stretchr/testify v1.9.0/go.mod h1:r2ic/lqez/lEtzL7wO/rwa5dbSLXVDPFyf8C91i36aY=
 golang.org/x/crypto v0.17.0 h1:r8bRNjWL3GshPW3gkd+RpvzWrZAwPS49OmTGZ/uhM4k=
 golang.org/x/crypto v0.17.0/go.mod h1:gCAAfMLgwOJRpTjQ2zCCt2OcSfYMTeZVSRtQlPC7Nq4=
-golang.org/x/sync v0.1.0 h1:wsuoTGHzEhffawBOhz5CYhcrV4IdKZbEyZjBMuTp12o=
-golang.org/x/sync v0.1.0/go.mod h1:RxMgew5VJxzue5/jJTE5uejpjVlOe/izrB70Jof72aM=
-golang.org/x/text v0.14.0 h1:ScX5w1eTa3QqT8oi6+ziP7dTV1S2+ALU0bI+0zXKWiQ=
-golang.org/x/text v0.14.0/go.mod h1:18ZOQIKpY8NJVqYksKHtTdi31H5itFRjB5/qKTNYzSU=
+golang.org/x/sync v0.22.0 h1:SZjpbeLmrCk4xhRSZFNZW5gFUeCeFgjekvI/+gfScek=
+golang.org/x/sync v0.22.0/go.mod h1:9xrNwdLfx4jkKbNva9FpL6vEN7evnE43NNNJQ2LF3+0=
+golang.org/x/sys v0.47.0 h1:o7XGOvZQCADBQQ4Y7VNq2dRWQR7JmOUW8Kxx4ZsNgWs=
+golang.org/x/sys v0.47.0/go.mod h1:4GL1E5IUh+htKOUEOaiffhrAeqysfVGipDYzABqnCmw=
+golang.org/x/text v0.40.0 h1:Ub2Z6/xjgF1WrYQz2nuITOEegKFtiIy+rieRJ5lHZKs=
+golang.org/x/text v0.40.0/go.mod h1:hpnzDAfGV753zIKo+wk3u1bVKCGPbrnF7+7LBF/UHVY=
+google.golang.org/protobuf v1.36.11 h1:fV6ZwhNocDyBLK0dj+fg8ektcVegBBuEolpbTQyBNVE=
+google.golang.org/protobuf v1.36.11/go.mod h1:HTf+CrKn2C3g5S8VImy6tdcUvCska2kB7j23XfzDpco=
 gopkg.in/check.v1 v0.0.0-20161208181325-20d25e280405/go.mod h1:Co6ibVJAznAaIkqp8huTwlJQCZ016jof/cbN4VW5Yz0=
 gopkg.in/natefinch/lumberjack.v2 v2.2.1 h1:bBRl1b0OH9s/DuPhuXpNl+VtCaJXFZ5/uEFST95x9zc=
 gopkg.in/natefinch/lumberjack.v2 v2.2.1/go.mod h1:YD8tP3GAjkrDg1eZH7EGmyESg/lsYskCTPBJVb9jqSc=

coordinator/internal/authctx/authctx.go

@@ -0,0 +1,43 @@
+// Package authctx carries the authenticated requester across the transport and
+// use-case layers without either one importing the other. The HTTP middleware
+// stamps a Requester after verifying a user's JWT; the job use cases read it to
+// record ownership and to enforce that a non-admin only touches their own jobs.
+package authctx
+
+import (
+	"context"
+
+	"github.com/google/uuid"
+)
+
+// Requester is the identity behind a request, derived from a verified JWT.
+// A request authenticated only by the shared worker/service token carries no
+// Requester at all (From returns ok=false), which is how worker traffic and
+// legacy unauthenticated-user traffic stay owner-less.
+type Requester struct {
+	UserID   uuid.UUID
+	Role     string
+	Verified bool
+}
+
+// IsAdmin reports whether the requester may act on any user's jobs.
+func (r Requester) IsAdmin() bool { return r.Role == "admin" }
+
+// IsTrusted reports whether workers this requester registers produce results
+// the coordinator accepts without quorum. Admins and verified contributors are
+// trusted; a plain unverified user is not.
+func (r Requester) IsTrusted() bool { return r.IsAdmin() || r.Verified }
+
+type ctxKey struct{}
+
+// With returns a copy of ctx carrying r.
+func With(ctx context.Context, r Requester) context.Context {
+	return context.WithValue(ctx, ctxKey{}, r)
+}
+
+// From returns the requester stamped by the middleware, or ok=false when the
+// request was not authenticated as a user.
+func From(ctx context.Context) (Requester, bool) {
+	r, ok := ctx.Value(ctxKey{}).(Requester)
+	return r, ok
+}

coordinator/internal/chunk/tsv.go

@@ -7,12 +7,19 @@ import (
 	"bytes"
 	"fmt"
 	"io"
+	"strings"
 )
 
 // ErrNoRows is returned when the input has a header but no data rows: a job with
 // zero tasks could never complete, so it is rejected at the source.
 var ErrNoRows = fmt.Errorf("input has no data rows")
 
+// maxShardBytes bounds the coordinator memory used by one in-progress shard.
+// The uploaded file may be much larger: it is first stored on disk, then split
+// in small bounded pieces. Operators can lower rowsPerShard when this limit is
+// reached rather than exhausting the coordinator process.
+const maxShardBytes = 64 << 20 // 64 MiB
+
 // SplitTSV reads a header-plus-rows text stream and cuts it into shards of at
 // most rowsPerShard data rows. Every shard repeats the header, so a worker can
 // parse its shard in isolation. emit is called once per shard, in order, with a
@@ -26,13 +33,25 @@ var ErrNoRows = fmt.Errorf("input has no data rows")
 // Only one shard is buffered at a time, so memory is bounded by shard size (a
 // worker-sized slice of the data), not by the size of the whole dataset.
 func SplitTSV(r io.Reader, rowsPerShard int, emit func(index int, shard io.Reader) error) error {
-	return SplitTSVLimit(r, rowsPerShard, 0, emit)
+	return splitTSVLimit(r, rowsPerShard, 0, nil, emit)
 }
 
 // SplitTSVLimit behaves like SplitTSV but emits no more than maxRows data rows.
 // A maxRows value of zero means unlimited. This lets an operator make a small,
 // representative pipeline check without materialising a second dataset file.
 func SplitTSVLimit(r io.Reader, rowsPerShard, maxRows int, emit func(index int, shard io.Reader) error) error {
+	return splitTSVLimit(r, rowsPerShard, maxRows, nil, emit)
+}
+
+// SplitChEMBLTSVLimit is the coordinator's scientific-upload splitter. It
+// validates the two columns every local SciMesh workload requires before any
+// shard task is persisted, while generic SplitTSV remains reusable for future
+// non-chemistry workloads.
+func SplitChEMBLTSVLimit(r io.Reader, rowsPerShard, maxRows int, emit func(index int, shard io.Reader) error) error {
+	return splitTSVLimit(r, rowsPerShard, maxRows, validateChEMBLHeader, emit)
+}
+
+func splitTSVLimit(r io.Reader, rowsPerShard, maxRows int, validateHeader func([]byte) error, emit func(index int, shard io.Reader) error) error {
 	if rowsPerShard <= 0 {
 		return fmt.Errorf("rowsPerShard must be positive, got %d", rowsPerShard)
 	}
@@ -51,6 +70,11 @@ func SplitTSVLimit(r io.Reader, rowsPerShard, maxRows int, emit func(index int,
 		return ErrNoRows // completely empty input
 	}
 	header := append([]byte(nil), sc.Bytes()...)
+	if validateHeader != nil {
+		if err := validateHeader(header); err != nil {
+			return err
+		}
+	}
 
 	var (
 		buf   bytes.Buffer
@@ -71,9 +95,15 @@ func SplitTSVLimit(r io.Reader, rowsPerShard, maxRows int, emit func(index int,
 
 	for sc.Scan() {
 		if rows == 0 {
+			if len(header)+1 > maxShardBytes {
+				return fmt.Errorf("TSV header exceeds maximum shard size of %d bytes", maxShardBytes)
+			}
 			buf.Write(header)
 			buf.WriteByte('\n')
 		}
+		if buf.Len()+len(sc.Bytes())+1 > maxShardBytes {
+			return fmt.Errorf("shard exceeds maximum size of %d bytes; lower rowsPerShard", maxShardBytes)
+		}
 		buf.Write(sc.Bytes())
 		buf.WriteByte('\n')
 		rows++
@@ -103,3 +133,17 @@ func SplitTSVLimit(r io.Reader, rowsPerShard, maxRows int, emit func(index int,
 	}
 	return nil
 }
+
+func validateChEMBLHeader(header []byte) error {
+	seen := make(map[string]struct{})
+	for _, field := range strings.Split(strings.TrimPrefix(string(header), "\ufeff"), "\t") {
+		seen[field] = struct{}{}
+	}
+	if _, ok := seen["chembl_id"]; !ok {
+		return fmt.Errorf("TSV is missing required column chembl_id")
+	}
+	if _, ok := seen["canonical_smiles"]; !ok {
+		return fmt.Errorf("TSV is missing required column canonical_smiles")
+	}
+	return nil
+}

coordinator/internal/chunk/tsv_test.go

@@ -119,6 +119,14 @@ func TestSplitLimitUsesOnlyLeadingDataRows(t *testing.T) {
 	}
 }
 
+func TestChEMBLSplitRejectsMissingRequiredColumns(t *testing.T) {
+	err := SplitChEMBLTSVLimit(strings.NewReader("id\tsmiles\nA\tCC\n"), 1, 0,
+		func(int, io.Reader) error { return nil })
+	if err == nil || !strings.Contains(err.Error(), "chembl_id") {
+		t.Errorf("err = %v, want missing-column error", err)
+	}
+}
+
 // The scanned bytes are reused by bufio; the shard buffer must copy them, or a
 // later row would corrupt an earlier one. This guards that copy.
 func TestSplitDoesNotAliasScannerBuffer(t *testing.T) {

coordinator/internal/domain/job.go

@@ -11,6 +11,7 @@ type JobStatus string
 const (
 	JobPending   JobStatus = "pending"
 	JobRunning   JobStatus = "running"
+	JobReducing  JobStatus = "reducing"
 	JobCompleted JobStatus = "completed"
 	JobFailed    JobStatus = "failed"
 	JobCancelled JobStatus = "cancelled"
@@ -19,13 +20,21 @@ const (
 // Job is one user submission that fans out into one or more tasks.
 type Job struct {
 	ID uuid.UUID
+	// OwnerID is the userservice user who submitted the job (JWT `sub`). nil
+	// when the job was created without user authentication. Not a foreign key:
+	// users live in a separate service/database.
+	OwnerID          *uuid.UUID
 	Workload         string
 	InputURI         string     // external input URI; empty for uploaded datasets
 	InputArtifactID  *uuid.UUID // uploaded input artifact; nil for URI submissions
+	ResultArtifactID *uuid.UUID
 	Parameters       map[string]any
 	Status           JobStatus
 	CreatedAt        time.Time
 	CompletedAt      *time.Time
+	ReducerStartedAt *time.Time
+	ErrorCode        *string
+	ErrorMessage     *string
 }
 
 // NewUploadedJob builds a job whose input was uploaded to the coordinator. The
@@ -114,6 +123,12 @@ func (p JobProgress) DeriveStatus() JobStatus {
 	switch {
 	case p.Job.Status == JobCancelled:
 		return JobCancelled
+	case p.Job.Status == JobFailed:
+		// A reducer may fail after every shard has completed. That terminal
+		// failure must not be overwritten by an otherwise-complete task count.
+		return JobFailed
+	case p.Job.Status == JobReducing:
+		return JobReducing
 	case p.Total == 0:
 		return JobPending
 	case p.Done == p.Total:

coordinator/internal/domain/job_test.go

@@ -82,6 +82,7 @@ func TestDeriveStatus(t *testing.T) {
 		{"done and failed", JobProgress{Total: 3, Done: 2, Failed: 1}, JobFailed},
 		{"failed but work remains", JobProgress{Total: 3, Pending: 1, Failed: 2}, JobRunning},
 		{"cancelled job wins over task histogram", JobProgress{Job: Job{Status: JobCancelled}, Total: 3, Done: 1, Cancelled: 2}, JobCancelled},
+		{"persisted reducer failure wins over completed tasks", JobProgress{Job: Job{Status: JobFailed}, Total: 3, Done: 3}, JobFailed},
 	}
 	for _, c := range cases {
 		t.Run(c.name, func(t *testing.T) {

coordinator/internal/domain/task.go

@@ -24,6 +24,10 @@ const (
 // ErrCodeLeaseExpired marks tasks failed by the reaper rather than by a worker.
 const ErrCodeLeaseExpired = "lease_expired"
 
+// ErrCodeQuorumFailed marks a task whose untrusted results never reached a
+// verifying quorum before its attempts ran out.
+const ErrCodeQuorumFailed = "quorum_failed"
+
 // Task is one independently executable chunk of a job.
 //
 // Nullable columns are pointers so "no lease" stays distinguishable from
@@ -216,6 +220,33 @@ func (t *Task) CompleteWith(resultArtifactID uuid.UUID, metrics map[string]any,
 	return nil
 }
 
+// ReleaseAfterVote returns an untrusted worker's task to the queue after its
+// result was recorded as a quorum vote but quorum was not yet reached, so a
+// different owner can compute it independently. When no attempts remain the task
+// fails: its untrusted results could not be verified.
+func (t *Task) ReleaseAfterVote(worker string, attempt int, now time.Time) error {
+	if t.Status == TaskCompleted {
+		return nil // settled by a concurrent quorum
+	}
+	if err := t.verifyLease(worker, attempt, now); err != nil {
+		return err
+	}
+	t.LeaseOwner = nil
+	t.LeaseExpiresAt = nil
+	t.Version++
+
+	if t.CanRetry() {
+		t.Status = TaskPending
+		return nil
+	}
+	code, msg := ErrCodeQuorumFailed, "untrusted results did not reach quorum"
+	t.ErrorCode = &code
+	t.ErrorMessage = &msg
+	t.Status = TaskFailed
+	t.CompletedAt = &now
+	return nil
+}
+
 // Fail records a worker-reported failure. A retryable failure with attempts
 // left returns the task to the queue; otherwise it terminates as failed.
 func (t *Task) Fail(worker string, attempt int, code, message string, retryable bool, now time.Time) error {

coordinator/internal/domain/worker.go

@@ -14,6 +14,17 @@ const (
 	WorkerOffline WorkerStatus = "offline"
 )
 
+// WorkerTrust says whether a worker's results are accepted directly or must
+// clear quorum cross-checking.
+type WorkerTrust string
+
+const (
+	// WorkerTrusted — lab machine (shared token) or a verified/admin contributor.
+	WorkerTrusted WorkerTrust = "trusted"
+	// WorkerUntrusted — a plain enthusiast; results are quarantined until quorum.
+	WorkerUntrusted WorkerTrust = "untrusted"
+)
+
 // Worker is a registered process/machine allowed to claim tasks. Its
 // capabilities are the allowlisted workload names it can run; the coordinator
 // never hands it a task outside that set.
@@ -22,6 +33,11 @@ type Worker struct {
 	Name         string
 	Capabilities []string
 	Status       WorkerStatus
+	// OwnerID is the userservice user who registered this worker; nil for a
+	// worker registered with the shared service token.
+	OwnerID *uuid.UUID
+	// TrustLevel decides whether this worker's results need quorum.
+	TrustLevel      WorkerTrust
 	LastHeartbeatAt time.Time
 	CreatedAt       time.Time
 	UpdatedAt       time.Time
@@ -29,6 +45,9 @@ type Worker struct {
 
 // NewWorker registers a worker. A worker with no capabilities could never be
 // handed a task, so an empty set is rejected rather than silently stored.
+//
+// Trust defaults to WorkerTrusted (the shared-token lab worker); the caller
+// overrides it for a volunteer registered through the userservice.
 func NewWorker(name string, capabilities []string, now time.Time) (*Worker, error) {
 	if len(capabilities) == 0 {
 		return nil, ErrInvalidInput
@@ -38,6 +57,7 @@ func NewWorker(name string, capabilities []string, now time.Time) (*Worker, erro
 		Name:            name,
 		Capabilities:    capabilities,
 		Status:          WorkerOnline,
+		TrustLevel:      WorkerTrusted,
 		LastHeartbeatAt: now,
 		CreatedAt:       now,
 		UpdatedAt:       now,

coordinator/internal/infra/config.go

@@ -26,6 +26,17 @@ type Config struct {
 	Token string
 	// Local operator UI credential. Empty disables the embedded UI entirely.
 	UIToken string
+	// Shared HS256 secret used to verify userservice-issued JWTs. When set, a
+	// submitter may authenticate with a JWT (in addition to workers using the
+	// shared token) and their jobs are stamped with owner_id. Empty disables
+	// user-JWT auth entirely — the pre-userservice behaviour. Must match the
+	// userservice's JWT_SECRET.
+	JWTSecret string
+	// Base URL of the userservice, e.g. http://userservice:8081. When set
+	// together with JWTSecret, the operator UI authenticates via userservice
+	// login/registration (cookie session) instead of the static UI_AUTH_TOKEN
+	// basic auth. Empty keeps the basic-auth UI.
+	UserserviceURL string
 
 	// Minimum log level: debug, info, warn, error.
 	LogLevel string
@@ -50,6 +61,9 @@ type Config struct {
 	LeaseDuration time.Duration
 	// Default attempt ceiling for newly created tasks.
 	DefaultMaxAttempts int
+	// How many distinct owners must agree on an untrusted result before it is
+	// accepted (trusted workers are accepted directly).
+	QuorumSize int
 	// How often the background lease-reaper runs.
 	ReaperInterval time.Duration
 	// A worker silent for longer than this is marked offline by the reaper.
@@ -80,6 +94,8 @@ func LoadConfig() (Config, error) {
 		// former name, still honoured so existing .env files keep working.
 		Token:              getEnv("COORDINATOR_TOKEN", os.Getenv("WORKER_AUTH_TOKEN")),
 		UIToken:            os.Getenv("UI_AUTH_TOKEN"),
+		JWTSecret:          os.Getenv("JWT_SECRET"),
+		UserserviceURL:     os.Getenv("USERSERVICE_URL"),
 		LogLevel:           getEnv("LOG_LEVEL", "info"),
 		LogFile:            os.Getenv("LOG_FILE"),
 		StorageDir:         getEnv("COORDINATOR_STORAGE_DIR", "./data"),
@@ -90,6 +106,7 @@ func LoadConfig() (Config, error) {
 		HeartbeatInterval:  15 * time.Second,
 		LeaseDuration:      2 * time.Minute,
 		DefaultMaxAttempts: 3,
+		QuorumSize:         2,
 		ReaperInterval:     30 * time.Second,
 		WorkerOfflineAfter: 1 * time.Minute,
 	}
@@ -100,6 +117,11 @@ func LoadConfig() (Config, error) {
 	if cfg.UIToken != "" && cfg.Token != "" && cfg.UIToken == cfg.Token {
 		return Config{}, fmt.Errorf("UI_AUTH_TOKEN must differ from the worker auth token")
 	}
+	// A short secret makes the HMAC brute-forceable; refuse a weak one rather
+	// than verify tokens against it.
+	if cfg.JWTSecret != "" && len(cfg.JWTSecret) < 32 {
+		return Config{}, fmt.Errorf("JWT_SECRET must be at least 32 bytes")
+	}
 
 	var err error
 	if cfg.DBMaxConns, err = getEnvInt32("DB_MAX_CONNS", cfg.DBMaxConns); err != nil {
@@ -129,6 +151,15 @@ func LoadConfig() (Config, error) {
 	if cfg.DefaultMaxAttempts, err = getEnvInt("DEFAULT_MAX_ATTEMPTS", cfg.DefaultMaxAttempts); err != nil {
 		return Config{}, err
 	}
+	if cfg.QuorumSize, err = getEnvInt("QUORUM_SIZE", cfg.QuorumSize); err != nil {
+		return Config{}, err
+	}
+	if cfg.QuorumSize < 1 {
+		return Config{}, fmt.Errorf("QUORUM_SIZE must be positive")
+	}
+	if cfg.DefaultMaxAttempts < 1 {
+		return Config{}, fmt.Errorf("DEFAULT_MAX_ATTEMPTS must be positive")
+	}
 
 	return cfg, nil
 }

coordinator/internal/infra/config_test.go

@@ -32,3 +32,14 @@ func TestLoadConfigAllowsDistinctUIAndWorkerTokens(t *testing.T) {
 		t.Fatalf("unexpected tokens: %+v", cfg)
 	}
 }
+
+func TestLoadConfigRejectsNonPositiveDefaultMaxAttempts(t *testing.T) {
+	t.Setenv("ENV_FILE", filepath.Join(t.TempDir(), "missing.env"))
+	t.Setenv("DATABASE_URL", "postgres://test")
+	t.Setenv("DEFAULT_MAX_ATTEMPTS", "0")
+
+	_, err := LoadConfig()
+	if err == nil || !strings.Contains(err.Error(), "DEFAULT_MAX_ATTEMPTS") {
+		t.Fatalf("LoadConfig error = %v, want default-attempt validation", err)
+	}
+}

coordinator/internal/memstore/memstore.go

@@ -160,17 +160,18 @@ func (r *TaskRepo) CancelByJob(_ context.Context, jobID uuid.UUID, now time.Time
 	return cancelled, nil
 }
 
-func (r *TaskRepo) ExpireLeases(ctx context.Context, now time.Time) (int64, error) {
+func (r *TaskRepo) ExpireLeases(ctx context.Context, now time.Time) ([]uuid.UUID, error) {
 	r.mu.Lock()
 	defer r.mu.Unlock()
-	var n int64
+	affected := make([]uuid.UUID, 0)
 	for _, t := range r.tasks {
-		if t.Status == domain.TaskLeased && t.LeaseExpiresAt != nil && t.LeaseExpiresAt.Before(now) {
+		if (t.Status == domain.TaskLeased || t.Status == domain.TaskRunning) &&
+			t.LeaseExpiresAt != nil && t.LeaseExpiresAt.Before(now) {
 			t.ExpireLease(now)
-			n++
+			affected = append(affected, t.JobID)
 		}
 	}
-	return n, nil
+	return affected, nil
 }
 
 // --- JobRepo -------------------------------------------------------------
@@ -215,6 +216,51 @@ func (r *JobRepo) UpdateStatus(ctx context.Context, id uuid.UUID, status domain.
 	return nil
 }
 
+func (r *JobRepo) ClaimReduction(_ context.Context, id uuid.UUID, startedAt time.Time) (bool, error) {
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	j, ok := r.jobs[id]
+	if !ok {
+		return false, domain.ErrJobNotFound
+	}
+	if j.Status != domain.JobReducing || j.ReducerStartedAt != nil {
+		return false, nil
+	}
+	j.ReducerStartedAt = &startedAt
+	return true, nil
+}
+
+func (r *JobRepo) CompleteWithResult(_ context.Context, id, resultArtifactID uuid.UUID, completedAt time.Time) error {
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	j, ok := r.jobs[id]
+	if !ok {
+		return domain.ErrJobNotFound
+	}
+	j.ResultArtifactID = &resultArtifactID
+	j.Status = domain.JobCompleted
+	j.CompletedAt = &completedAt
+	j.ReducerStartedAt = nil
+	j.ErrorCode = nil
+	j.ErrorMessage = nil
+	return nil
+}
+
+func (r *JobRepo) FailReduction(_ context.Context, id uuid.UUID, code, message string, completedAt time.Time) error {
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	j, ok := r.jobs[id]
+	if !ok {
+		return domain.ErrJobNotFound
+	}
+	j.Status = domain.JobFailed
+	j.CompletedAt = &completedAt
+	j.ReducerStartedAt = nil
+	j.ErrorCode = &code
+	j.ErrorMessage = &message
+	return nil
+}
+
 // --- WorkerRepo ----------------------------------------------------------
 
 type WorkerRepo struct {
@@ -298,6 +344,23 @@ func (r *ArtifactRepo) Get(ctx context.Context, id uuid.UUID) (*domain.Artifact,
 	return &cp, nil
 }
 
+func (r *ArtifactRepo) FindPartialResult(_ context.Context, taskID uuid.UUID, attempt int) (*domain.Artifact, error) {
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	for _, a := range r.arts {
+		if a.TaskID != nil && *a.TaskID == taskID && a.Kind == domain.ArtifactPartialResult &&
+			a.Attempt != nil && *a.Attempt == attempt {
+			return cloneArtifact(a), nil
+		}
+	}
+	return nil, nil
+}
+
+func cloneArtifact(a *domain.Artifact) *domain.Artifact {
+	cp := *a
+	return &cp
+}
+
 // --- BlobStore -----------------------------------------------------------
 
 type BlobStore struct {
@@ -354,3 +417,36 @@ func contains(ss []string, s string) bool {
 	}
 	return false
 }
+
+// TaskResultRepo is an in-memory usecase.TaskResultRepository: one vote per
+// (task, owner).
+type TaskResultRepo struct {
+	mu    sync.Mutex
+	votes map[uuid.UUID]map[uuid.UUID]string // taskID -> ownerID -> sha256
+}
+
+func NewTaskResultRepo() *TaskResultRepo {
+	return &TaskResultRepo{votes: make(map[uuid.UUID]map[uuid.UUID]string)}
+}
+
+func (r *TaskResultRepo) RecordVote(_ context.Context, taskID, ownerID uuid.UUID, sha256 string, _ uuid.UUID) error {
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	if r.votes[taskID] == nil {
+		r.votes[taskID] = make(map[uuid.UUID]string)
+	}
+	r.votes[taskID][ownerID] = sha256
+	return nil
+}
+
+func (r *TaskResultRepo) CountAgreeing(_ context.Context, taskID uuid.UUID, sha256 string) (int, error) {
+	r.mu.Lock()
+	defer r.mu.Unlock()
+	n := 0
+	for _, s := range r.votes[taskID] {
+		if s == sha256 {
+			n++
+		}
+	}
+	return n, nil
+}

coordinator/internal/memstore/ui_read.go

@@ -27,7 +27,7 @@ var _ usecase.UIReadRepository = (*UIReadRepo)(nil)
 func (r *UIReadRepo) GetJob(ctx context.Context, id uuid.UUID) (*domain.Job, error) {
 	return r.jobs.Get(ctx, id)
 }
-func (r *UIReadRepo) ListJobs(_ context.Context, limit int) ([]domain.Job, error) {
+func (r *UIReadRepo) ListJobs(_ context.Context, owner *uuid.UUID, limit int) ([]domain.Job, error) {
 	if limit < 1 || limit > 100 {
 		return nil, domain.ErrInvalidInput
 	}
@@ -35,6 +35,9 @@ func (r *UIReadRepo) ListJobs(_ context.Context, limit int) ([]domain.Job, error
 	defer r.jobs.mu.Unlock()
 	out := make([]domain.Job, 0, len(r.jobs.jobs))
 	for _, job := range r.jobs.jobs {
+		if owner != nil && (job.OwnerID == nil || *job.OwnerID != *owner) {
+			continue
+		}
 		out = append(out, *job)
 	}
 	sort.Slice(out, func(i, j int) bool {
@@ -60,6 +63,18 @@ func (r *UIReadRepo) ListTasksByJob(_ context.Context, jobID uuid.UUID) ([]domai
 	sort.Slice(out, func(i, j int) bool { return out[i].ChunkIndex < out[j].ChunkIndex })
 	return out, nil
 }
+
+func (r *UIReadRepo) ListTasksByJobs(ctx context.Context, jobIDs []uuid.UUID) (map[uuid.UUID][]domain.Task, error) {
+	out := make(map[uuid.UUID][]domain.Task, len(jobIDs))
+	for _, id := range jobIDs {
+		tasks, err := r.ListTasksByJob(ctx, id)
+		if err != nil {
+			return nil, err
+		}
+		out[id] = tasks
+	}
+	return out, nil
+}
 func (r *UIReadRepo) ListWorkers(_ context.Context, limit int) ([]domain.Worker, error) {
 	if limit < 1 || limit > 100 {
 		return nil, domain.ErrInvalidInput

coordinator/internal/metrics/business.go

@@ -0,0 +1,66 @@
+package metrics
+
+import (
+	"context"
+	"time"
+
+	"github.com/prometheus/client_golang/prometheus"
+)
+
+// Stats is a point-in-time snapshot of the coordinator's domain state: counts of
+// tasks, jobs, and workers keyed by their status. Maps are expected to be
+// zero-filled by the provider so every known status is always present, giving
+// the dashboard flat zero lines instead of gaps.
+type Stats struct {
+	Tasks   map[string]int
+	Jobs    map[string]int
+	Workers map[string]int
+}
+
+// StatsFunc returns the current snapshot. It is called on every scrape, so it
+// must be a cheap aggregate query.
+type StatsFunc func(context.Context) (Stats, error)
+
+// RegisterBusiness registers a collector that reports domain-state gauges
+// (scimesh_tasks/jobs/workers by status) sourced from collect on each scrape.
+// Deriving the gauges at scrape time keeps them fresh without a background
+// goroutine, and a failed query simply yields no samples for that scrape.
+func (m *Metrics) RegisterBusiness(collect StatsFunc) {
+	m.reg.MustRegister(&businessCollector{
+		collect: collect,
+		tasks:   prometheus.NewDesc("scimesh_tasks", "Tasks by status.", []string{"status"}, nil),
+		jobs:    prometheus.NewDesc("scimesh_jobs", "Jobs by status.", []string{"status"}, nil),
+		workers: prometheus.NewDesc("scimesh_workers", "Workers by status.", []string{"status"}, nil),
+	})
+}
+
+type businessCollector struct {
+	collect              StatsFunc
+	tasks, jobs, workers *prometheus.Desc
+}
+
+func (c *businessCollector) Describe(ch chan<- *prometheus.Desc) {
+	ch <- c.tasks
+	ch <- c.jobs
+	ch <- c.workers
+}
+
+func (c *businessCollector) Collect(ch chan<- prometheus.Metric) {
+	// A bounded query so one slow scrape cannot stall Prometheus.
+	ctx, cancel := context.WithTimeout(context.Background(), 3*time.Second)
+	defer cancel()
+
+	s, err := c.collect(ctx)
+	if err != nil {
+		return // no samples this scrape; Prometheus keeps the last value
+	}
+	emit(ch, c.tasks, s.Tasks)
+	emit(ch, c.jobs, s.Jobs)
+	emit(ch, c.workers, s.Workers)
+}
+
+func emit(ch chan<- prometheus.Metric, desc *prometheus.Desc, counts map[string]int) {
+	for status, n := range counts {
+		ch <- prometheus.MustNewConstMetric(desc, prometheus.GaugeValue, float64(n), status)
+	}
+}

coordinator/internal/metrics/business_test.go

@@ -0,0 +1,51 @@
+package metrics
+
+import (
+	"context"
+	"errors"
+	"net/http"
+	"net/http/httptest"
+	"strings"
+	"testing"
+)
+
+func scrape(t *testing.T, m *Metrics) string {
+	t.Helper()
+	rec := httptest.NewRecorder()
+	req, _ := http.NewRequestWithContext(context.Background(), http.MethodGet, "/metrics", nil)
+	m.Handler().ServeHTTP(rec, req)
+	return rec.Body.String()
+}
+
+func TestBusinessCollectorEmitsGauges(t *testing.T) {
+	m := New()
+	m.RegisterBusiness(func(context.Context) (Stats, error) {
+		return Stats{
+			Tasks:   map[string]int{"pending": 3, "running": 1, "completed": 0},
+			Jobs:    map[string]int{"running": 2},
+			Workers: map[string]int{"online": 4},
+		}, nil
+	})
+
+	body := scrape(t, m)
+	for _, want := range []string{
+		`scimesh_tasks{status="pending"} 3`,
+		`scimesh_tasks{status="completed"} 0`,
+		`scimesh_jobs{status="running"} 2`,
+		`scimesh_workers{status="online"} 4`,
+	} {
+		if !strings.Contains(body, want) {
+			t.Errorf("metrics missing %q\n%s", want, body)
+		}
+	}
+}
+
+func TestBusinessCollectorSkipsOnError(t *testing.T) {
+	m := New()
+	m.RegisterBusiness(func(context.Context) (Stats, error) {
+		return Stats{}, errors.New("db down")
+	})
+	if strings.Contains(scrape(t, m), "scimesh_tasks") {
+		t.Error("a failed snapshot must emit no business samples")
+	}
+}

coordinator/internal/metrics/metrics.go

@@ -0,0 +1,112 @@
+// Package metrics exposes Prometheus instrumentation for the coordinator: an
+// HTTP RED middleware (rate, errors, duration) plus the standard Go runtime and
+// process collectors, all on a private registry so nothing leaks in from global
+// state.
+package metrics
+
+import (
+	"net/http"
+	"regexp"
+	"strconv"
+	"strings"
+	"time"
+
+	"github.com/prometheus/client_golang/prometheus"
+	"github.com/prometheus/client_golang/prometheus/collectors"
+	"github.com/prometheus/client_golang/prometheus/promhttp"
+)
+
+type Metrics struct {
+	reg      *prometheus.Registry
+	requests *prometheus.CounterVec
+	duration *prometheus.HistogramVec
+}
+
+// New builds the registry and registers the runtime, process, and HTTP metrics.
+func New() *Metrics {
+	reg := prometheus.NewRegistry()
+	reg.MustRegister(
+		collectors.NewGoCollector(),
+		collectors.NewProcessCollector(collectors.ProcessCollectorOpts{}),
+	)
+
+	requests := prometheus.NewCounterVec(prometheus.CounterOpts{
+		Namespace: "scimesh",
+		Subsystem: "http",
+		Name:      "requests_total",
+		Help:      "HTTP requests, labelled by method, normalized route, and status.",
+	}, []string{"method", "route", "status"})
+
+	duration := prometheus.NewHistogramVec(prometheus.HistogramOpts{
+		Namespace: "scimesh",
+		Subsystem: "http",
+		Name:      "request_duration_seconds",
+		Help:      "HTTP request duration in seconds.",
+		Buckets:   prometheus.DefBuckets,
+	}, []string{"method", "route"})
+
+	reg.MustRegister(requests, duration)
+	return &Metrics{reg: reg, requests: requests, duration: duration}
+}
+
+// Handler serves the metrics in Prometheus text format.
+func (m *Metrics) Handler() http.Handler {
+	return promhttp.HandlerFor(m.reg, promhttp.HandlerOpts{})
+}
+
+// Registry exposes the registry so callers can register extra collectors.
+func (m *Metrics) Registry() *prometheus.Registry { return m.reg }
+
+// Middleware records one request into the RED metrics. It normalizes the path
+// so per-id routes collapse to a single low-cardinality label.
+func (m *Metrics) Middleware(next http.Handler) http.Handler {
+	return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		start := time.Now()
+		rec := &statusRecorder{ResponseWriter: w, status: http.StatusOK}
+		next.ServeHTTP(rec, r)
+
+		route := normalizeRoute(r.URL.Path)
+		m.requests.WithLabelValues(r.Method, route, strconv.Itoa(rec.status)).Inc()
+		m.duration.WithLabelValues(r.Method, route).Observe(time.Since(start).Seconds())
+	})
+}
+
+type statusRecorder struct {
+	http.ResponseWriter
+	status int
+}
+
+func (s *statusRecorder) WriteHeader(code int) {
+	s.status = code
+	s.ResponseWriter.WriteHeader(code)
+}
+
+var uuidRe = regexp.MustCompile(`^[0-9a-fA-F]{8}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{4}-[0-9a-fA-F]{12}$`)
+
+// normalizeRoute collapses uuid and numeric path segments to {id}, keeping the
+// route label cardinality bounded (otherwise every job/task id would be its own
+// time series).
+func normalizeRoute(path string) string {
+	if path == "" {
+		return "/"
+	}
+	segs := strings.Split(path, "/")
+	for i, s := range segs {
+		if s == "" {
+			continue
+		}
+		if uuidRe.MatchString(s) || isAllDigits(s) {
+			segs[i] = "{id}"
+		}
+	}
+	return strings.Join(segs, "/")
+}
+
+func isAllDigits(s string) bool {
+	for _, r := range s {
+		if r < '0' || r > '9' {
+			return false
+		}
+	}
+	return s != ""
+}

coordinator/internal/metrics/metrics_test.go

@@ -0,0 +1,47 @@
+package metrics
+
+import (
+	"context"
+	"net/http"
+	"net/http/httptest"
+	"strings"
+	"testing"
+)
+
+func TestNormalizeRoute(t *testing.T) {
+	cases := map[string]string{
+		"/health": "/health",
+		"/jobs/3f2504e0-4f89-41d3-9a0c-0305e82c3301":         "/jobs/{id}",
+		"/tasks/3f2504e0-4f89-41d3-9a0c-0305e82c3301/result": "/tasks/{id}/result",
+		"/ui/jobs/12345": "/ui/jobs/{id}",
+		"/":              "/",
+	}
+	for in, want := range cases {
+		if got := normalizeRoute(in); got != want {
+			t.Errorf("normalizeRoute(%q) = %q, want %q", in, got, want)
+		}
+	}
+}
+
+func TestMiddlewareAndHandler(t *testing.T) {
+	m := New()
+	h := m.Middleware(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
+		w.WriteHeader(http.StatusCreated)
+	}))
+
+	req, _ := http.NewRequestWithContext(context.Background(), http.MethodPost, "/jobs/3f2504e0-4f89-41d3-9a0c-0305e82c3301", nil)
+	h.ServeHTTP(httptest.NewRecorder(), req)
+
+	// Scrape and confirm the request was recorded under the normalized route.
+	rec := httptest.NewRecorder()
+	greq, _ := http.NewRequestWithContext(context.Background(), http.MethodGet, "/metrics", nil)
+	m.Handler().ServeHTTP(rec, greq)
+
+	body := rec.Body.String()
+	if !strings.Contains(body, `scimesh_http_requests_total{method="POST",route="/jobs/{id}",status="201"}`) {
+		t.Errorf("requests_total not recorded as expected; body:\n%s", body)
+	}
+	if !strings.Contains(body, "go_goroutines") {
+		t.Error("Go runtime collector not registered")
+	}
+}

coordinator/internal/reducer/similarity_search.go

@@ -0,0 +1,195 @@
+// Package reducer contains deterministic, coordinator-side result reductions.
+package reducer
+
+import (
+	"bytes"
+	"encoding/csv"
+	"errors"
+	"fmt"
+	"io"
+	"math"
+	"sort"
+	"strconv"
+)
+
+var searchHeader = []string{"rank", "chembl_id", "canonical_smiles", "similarity"}
+
+type similarityMatch struct {
+	similarity float64
+	id         string
+	smiles     string
+}
+
+// ReduceSimilaritySearch streams worker-local top-k CSVs into the exact global
+// top-k. Each partial is validated before it can affect the final artifact.
+func ReduceSimilaritySearch(partials []io.Reader, parameters map[string]any) ([]byte, error) {
+	topK, err := positiveInt(parameters["top_k"], 20)
+	if err != nil {
+		return nil, err
+	}
+	direction, err := thresholdDirection(parameters["threshold_direction"])
+	if err != nil {
+		return nil, err
+	}
+	h := &matchHeap{direction: direction}
+	for _, partial := range partials {
+		if err := readPartial(partial, direction, func(match similarityMatch) {
+			if len(h.items) < topK {
+				heapPush(h, match)
+				return
+			}
+			if better(match, h.items[0], direction) {
+				h.items[0] = match
+				heapDown(h, 0)
+			}
+		}); err != nil {
+			return nil, err
+		}
+	}
+
+	matches := append([]similarityMatch(nil), h.items...)
+	sort.Slice(matches, func(i, j int) bool { return better(matches[i], matches[j], direction) })
+	var out bytes.Buffer
+	writer := csv.NewWriter(&out)
+	if err := writer.Write(searchHeader); err != nil {
+		return nil, err
+	}
+	for index, match := range matches {
+		if err := writer.Write([]string{
+			strconv.Itoa(index + 1), match.id, match.smiles, fmt.Sprintf("%.6f", match.similarity),
+		}); err != nil {
+			return nil, err
+		}
+	}
+	writer.Flush()
+	if err := writer.Error(); err != nil {
+		return nil, err
+	}
+	return out.Bytes(), nil
+}
+
+func readPartial(input io.Reader, direction string, consume func(similarityMatch)) error {
+	reader := csv.NewReader(input)
+	header, err := reader.Read()
+	if err != nil {
+		return fmt.Errorf("read partial header: %w", err)
+	}
+	if !equalStrings(header, searchHeader) {
+		return fmt.Errorf("partial result has an invalid CSV header")
+	}
+	var previous *similarityMatch
+	for rank := 1; ; rank++ {
+		row, err := reader.Read()
+		if errors.Is(err, io.EOF) {
+			return nil
+		}
+		if err != nil {
+			return fmt.Errorf("read partial row: %w", err)
+		}
+		if len(row) != len(searchHeader) || row[0] != strconv.Itoa(rank) {
+			return fmt.Errorf("partial result has an invalid rank")
+		}
+		score, err := strconv.ParseFloat(row[3], 64)
+		if err != nil || math.IsNaN(score) || math.IsInf(score, 0) || score < 0 || score > 1 {
+			return fmt.Errorf("partial result has an invalid similarity")
+		}
+		match := similarityMatch{similarity: score, id: row[1], smiles: row[2]}
+		if previous != nil && better(match, *previous, direction) {
+			return fmt.Errorf("partial result is not sorted deterministically")
+		}
+		previous = &match
+		consume(match)
+	}
+}
+
+func positiveInt(value any, fallback int) (int, error) {
+	if value == nil {
+		return fallback, nil
+	}
+	switch n := value.(type) {
+	case int:
+		if n > 0 {
+			return n, nil
+		}
+	case int64:
+		if n > 0 && n <= math.MaxInt {
+			return int(n), nil
+		}
+	case float64:
+		if n > 0 && n == math.Trunc(n) && n <= math.MaxInt {
+			return int(n), nil
+		}
+	}
+	return 0, fmt.Errorf("top_k must be a positive integer")
+}
+
+func thresholdDirection(value any) (string, error) {
+	if value == nil {
+		return "greater", nil
+	}
+	direction, ok := value.(string)
+	if !ok || (direction != "greater" && direction != "less") {
+		return "", fmt.Errorf("threshold_direction must be greater or less")
+	}
+	return direction, nil
+}
+
+func better(left, right similarityMatch, direction string) bool {
+	if left.similarity != right.similarity {
+		if direction == "less" {
+			return left.similarity < right.similarity
+		}
+		return left.similarity > right.similarity
+	}
+	if left.id != right.id {
+		return left.id < right.id
+	}
+	return left.smiles < right.smiles
+}
+
+func equalStrings(left, right []string) bool {
+	if len(left) != len(right) {
+		return false
+	}
+	for index := range left {
+		if left[index] != right[index] {
+			return false
+		}
+	}
+	return true
+}
+
+// matchHeap keeps the worst retained match at index zero.
+type matchHeap struct {
+	items     []similarityMatch
+	direction string
+}
+
+func heapPush(h *matchHeap, value similarityMatch) {
+	h.items = append(h.items, value)
+	for child := len(h.items) - 1; child > 0; {
+		parent := (child - 1) / 2
+		if !better(h.items[parent], h.items[child], h.direction) {
+			break
+		}
+		h.items[parent], h.items[child] = h.items[child], h.items[parent]
+		child = parent
+	}
+}
+
+func heapDown(h *matchHeap, parent int) {
+	for {
+		child := parent*2 + 1
+		if child >= len(h.items) {
+			return
+		}
+		if right := child + 1; right < len(h.items) && better(h.items[child], h.items[right], h.direction) {
+			child = right
+		}
+		if !better(h.items[parent], h.items[child], h.direction) {
+			return
+		}
+		h.items[parent], h.items[child] = h.items[child], h.items[parent]
+		parent = child
+	}
+}

coordinator/internal/reducer/similarity_search_test.go

@@ -0,0 +1,41 @@
+package reducer
+
+import (
+	"io"
+	"strings"
+	"testing"
+)
+
+func TestReduceSimilaritySearchKeepsExactCrossShardRanking(t *testing.T) {
+	first := strings.NewReader("rank,chembl_id,canonical_smiles,similarity\n1,B,CCC,0.50000048\n2,C,CCCC,0.1\n")
+	second := strings.NewReader("rank,chembl_id,canonical_smiles,similarity\n1,A,CC,0.50000049\n")
+
+	output, err := ReduceSimilaritySearch([]io.Reader{first, second}, map[string]any{"top_k": 2})
+	if err != nil {
+		t.Fatal(err)
+	}
+	want := "rank,chembl_id,canonical_smiles,similarity\n1,A,CC,0.500000\n2,B,CCC,0.500000\n"
+	if string(output) != want {
+		t.Fatalf("output = %q, want %q", output, want)
+	}
+}
+
+func TestReduceSimilaritySearchSupportsLeastSimilarDirection(t *testing.T) {
+	partial := strings.NewReader("rank,chembl_id,canonical_smiles,similarity\n1,A,CC,0.1\n2,B,CCC,0.8\n")
+	output, err := ReduceSimilaritySearch([]io.Reader{partial}, map[string]any{
+		"top_k": 1, "threshold_direction": "less",
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if got, want := string(output), "rank,chembl_id,canonical_smiles,similarity\n1,A,CC,0.100000\n"; got != want {
+		t.Fatalf("output = %q, want %q", got, want)
+	}
+}
+
+func TestReduceSimilaritySearchRejectsMalformedPartial(t *testing.T) {
+	partial := strings.NewReader("rank,chembl_id,canonical_smiles,similarity\n2,A,CC,0.1\n")
+	if _, err := ReduceSimilaritySearch([]io.Reader{partial}, nil); err == nil {
+		t.Fatal("expected malformed rank error")
+	}
+}

coordinator/internal/storage/postgres/artifact_repo.go

@@ -70,3 +70,33 @@ func (r *ArtifactRepo) Get(ctx context.Context, id uuid.UUID) (*domain.Artifact,
 	a.Kind = domain.ArtifactKind(kind)
 	return &a, nil
 }
+
+func (r *ArtifactRepo) FindPartialResult(ctx context.Context, taskID uuid.UUID, attempt int) (*domain.Artifact, error) {
+	sql, args, err := psql.Select(artifactColumns...).
+		From("artifacts").
+		Where(sq.Eq{
+			"task_id": taskID,
+			"attempt": attempt,
+			"kind":    string(domain.ArtifactPartialResult),
+		}).
+		ToSql()
+	if err != nil {
+		return nil, err
+	}
+
+	var (
+		a    domain.Artifact
+		kind string
+	)
+	err = conn(ctx, r.pool).QueryRow(ctx, sql, args...).Scan(
+		&a.ID, &a.JobID, &a.TaskID, &a.Attempt, &kind, &a.Filename, &a.StorageKey,
+		&a.ContentType, &a.SizeBytes, &a.SHA256, &a.CreatedAt)
+	if errors.Is(err, pgx.ErrNoRows) {
+		return nil, nil
+	}
+	if err != nil {
+		return nil, fmt.Errorf("find partial result: %w", err)
+	}
+	a.Kind = domain.ArtifactKind(kind)
+	return &a, nil
+}

coordinator/internal/storage/postgres/integration_test.go

@@ -90,6 +90,76 @@ func TestCreateJobPersistsEveryTask(t *testing.T) {
 	}
 }
 
+func TestClaimReductionIsAtomic(t *testing.T) {
+	pool := testPool(t)
+	job, _ := seedJob(t, pool, 1)
+	repo := NewJobRepo(pool)
+	ctx := context.Background()
+	if err := repo.UpdateStatus(ctx, job.ID, domain.JobReducing, nil); err != nil {
+		t.Fatal(err)
+	}
+
+	var (
+		wg      sync.WaitGroup
+		mu      sync.Mutex
+		claimed int
+	)
+	for range 8 {
+		wg.Add(1)
+		go func() {
+			defer wg.Done()
+			ok, err := repo.ClaimReduction(context.Background(), job.ID, time.Now().UTC())
+			if err != nil {
+				t.Errorf("claim reduction: %v", err)
+				return
+			}
+			if ok {
+				mu.Lock()
+				claimed++
+				mu.Unlock()
+			}
+		}()
+	}
+	wg.Wait()
+	if claimed != 1 {
+		t.Fatalf("reducer claims = %d, want 1", claimed)
+	}
+	stored, err := repo.Get(ctx, job.ID)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if stored.Status != domain.JobReducing || stored.ReducerStartedAt == nil {
+		t.Fatalf("stored reduction state = %+v", stored)
+	}
+}
+
+func TestUIReadRepoListsReducerFields(t *testing.T) {
+	pool := testPool(t)
+	job, _ := seedJob(t, pool, 1)
+	jobs := NewJobRepo(pool)
+	ctx := context.Background()
+	if err := jobs.UpdateStatus(ctx, job.ID, domain.JobReducing, nil); err != nil {
+		t.Fatal(err)
+	}
+	if claimed, err := jobs.ClaimReduction(ctx, job.ID, time.Now().UTC()); err != nil || !claimed {
+		t.Fatalf("claim reduction = (%v, %v)", claimed, err)
+	}
+	listed, err := NewUIReadRepo(pool).ListJobs(ctx, nil, 20)
+	if err != nil {
+		t.Fatalf("list UI jobs: %v", err)
+	}
+	for _, item := range listed {
+		if item.ID != job.ID {
+			continue
+		}
+		if item.Status != domain.JobReducing || item.ReducerStartedAt == nil {
+			t.Fatalf("UI reducer projection = %+v", item)
+		}
+		return
+	}
+	t.Fatalf("seeded job %s is missing from UI list", job.ID)
+}
+
 // A job must land whole or not at all: a half-created job leaves chunks no
 // worker could ever complete.
 func TestCreateJobRollsBackOnFailure(t *testing.T) {
@@ -336,8 +406,9 @@ func TestCompleteTaskReplayIsIdempotent(t *testing.T) {
 	job, _ := seedJob(t, pool, 1)
 
 	tasks, jobs, artifacts, tx := NewTaskRepo(pool), NewJobRepo(pool), NewArtifactRepo(pool), NewTxManager(pool)
+	workers, results := NewWorkerRepo(pool), NewTaskResultRepo(pool)
 	clk := fixedClock{now: time.Now().UTC()}
-	uc := usecase.NewCompleteTask(tasks, jobs, artifacts, tx, clk)
+	uc := usecase.NewCompleteTask(tasks, jobs, artifacts, workers, results, tx, clk, 2)
 
 	claimed, err := tasks.ClaimNext(ctx, usecase.ClaimFilter{
 		Owner: "worker-1", Now: clk.now, LeaseUntil: clk.now.Add(time.Minute),
@@ -360,12 +431,23 @@ func TestCompleteTaskReplayIsIdempotent(t *testing.T) {
 		t.Errorf("replay must be idempotent, got %v", err)
 	}
 
-	// A different result artifact for the same task is a genuine conflict.
-	art2 := seedArtifact(t, pool, job.ID, &claimed.ID, domain.ArtifactPartialResult)
-	other := in
-	other.ResultArtifactID = art2.ID
-	if _, err := uc.Execute(ctx, other); !errors.Is(err, domain.ErrResultConflict) {
-		t.Errorf("err = %v, want ErrResultConflict", err)
+}
+
+func TestPartialResultIsUniquePerTaskAttempt(t *testing.T) {
+	pool := testPool(t)
+	ctx := context.Background()
+	job, tasks := seedJob(t, pool, 1)
+	taskID := tasks[0].ID
+	first := seedArtifact(t, pool, job.ID, &taskID, domain.ArtifactPartialResult)
+	second, err := domain.NewArtifact(job.ID, &taskID, domain.ArtifactPartialResult, "retry.csv", "text/csv", time.Now().UTC())
+	if err != nil {
+		t.Fatal(err)
+	}
+	attempt := 1
+	second.Attempt = &attempt
+	second.SetContent("other-sha", 5)
+	if err := NewArtifactRepo(pool).Insert(ctx, second); err == nil {
+		t.Fatalf("second partial artifact for %s/%d was accepted after %s", taskID, attempt, first.ID)
 	}
 }
 

coordinator/internal/storage/postgres/job_repo.go

@@ -25,14 +25,18 @@ func NewJobRepo(pool *pgxpool.Pool) *JobRepo {
 
 var _ usecase.JobRepository = (*JobRepo)(nil)
 
-var jobColumns = []string{"id", "workload", "input_uri", "parameters", "status", "created_at", "completed_at"}
+var jobColumns = []string{
+	"id", "workload", "input_uri", "parameters", "status", "created_at", "completed_at",
+	"input_artifact_id", "result_artifact_id", "error_code", "error_message", "reducer_started_at",
+	"owner_id",
+}
 
 // Insert runs inside the caller's transaction, alongside the job's tasks — that
 // is what makes "all tasks or none" hold.
 func (r *JobRepo) Insert(ctx context.Context, j *domain.Job) error {
 	sql, args, err := psql.Insert("jobs").
-		Columns("id", "workload", "input_uri", "parameters", "status", "created_at").
-		Values(j.ID, j.Workload, j.InputURI, jsonbOrEmpty(j.Parameters), string(j.Status), j.CreatedAt).
+		Columns("id", "workload", "input_uri", "parameters", "status", "created_at", "owner_id").
+		Values(j.ID, j.Workload, j.InputURI, jsonbOrEmpty(j.Parameters), string(j.Status), j.CreatedAt, j.OwnerID).
 		ToSql()
 	if err != nil {
 		return err
@@ -55,7 +59,9 @@ func (r *JobRepo) Get(ctx context.Context, id uuid.UUID) (*domain.Job, error) {
 		status string
 	)
 	err = conn(ctx, r.pool).QueryRow(ctx, sql, args...).Scan(
-		&j.ID, &j.Workload, &j.InputURI, &j.Parameters, &status, &j.CreatedAt, &j.CompletedAt)
+		&j.ID, &j.Workload, &j.InputURI, &j.Parameters, &status, &j.CreatedAt, &j.CompletedAt,
+		&j.InputArtifactID, &j.ResultArtifactID, &j.ErrorCode, &j.ErrorMessage, &j.ReducerStartedAt,
+		&j.OwnerID)
 	if errors.Is(err, pgx.ErrNoRows) {
 		return nil, domain.ErrJobNotFound
 	}
@@ -66,6 +72,64 @@ func (r *JobRepo) Get(ctx context.Context, id uuid.UUID) (*domain.Job, error) {
 	return &j, nil
 }
 
+func (r *JobRepo) ClaimReduction(ctx context.Context, id uuid.UUID, startedAt time.Time) (bool, error) {
+	sql, args, err := psql.Update("jobs").
+		Set("reducer_started_at", startedAt).
+		Where(sq.Eq{"id": id, "status": string(domain.JobReducing), "reducer_started_at": nil}).
+		ToSql()
+	if err != nil {
+		return false, err
+	}
+	tag, err := conn(ctx, r.pool).Exec(ctx, sql, args...)
+	if err != nil {
+		return false, err
+	}
+	return tag.RowsAffected() == 1, nil
+}
+
+func (r *JobRepo) CompleteWithResult(ctx context.Context, id, resultArtifactID uuid.UUID, completedAt time.Time) error {
+	sql, args, err := psql.Update("jobs").
+		SetMap(map[string]any{
+			"status":             string(domain.JobCompleted),
+			"result_artifact_id": resultArtifactID,
+			"completed_at":       completedAt,
+			"reducer_started_at": nil,
+			"error_code":         nil,
+			"error_message":      nil,
+		}).
+		Where(sq.Eq{"id": id, "status": string(domain.JobReducing)}).
+		ToSql()
+	if err != nil {
+		return err
+	}
+	tag, err := conn(ctx, r.pool).Exec(ctx, sql, args...)
+	if err != nil {
+		return err
+	}
+	if tag.RowsAffected() == 0 {
+		return domain.ErrJobNotFound
+	}
+	return nil
+}
+
+func (r *JobRepo) FailReduction(ctx context.Context, id uuid.UUID, code, message string, completedAt time.Time) error {
+	sql, args, err := psql.Update("jobs").
+		SetMap(map[string]any{
+			"status":             string(domain.JobFailed),
+			"completed_at":       completedAt,
+			"error_code":         code,
+			"error_message":      message,
+			"reducer_started_at": nil,
+		}).
+		Where(sq.Eq{"id": id, "status": string(domain.JobReducing)}).
+		ToSql()
+	if err != nil {
+		return err
+	}
+	_, err = conn(ctx, r.pool).Exec(ctx, sql, args...)
+	return err
+}
+
 func (r *JobRepo) UpdateStatus(ctx context.Context, id uuid.UUID,
 	status domain.JobStatus, completedAt *time.Time) error {
 

coordinator/internal/storage/postgres/stats_repo.go

@@ -0,0 +1,65 @@
+package postgres
+
+import (
+	"context"
+	"fmt"
+
+	"github.com/jackc/pgx/v5/pgxpool"
+
+	"github.com/emil28092005/SciMesh/coordinator/internal/domain"
+)
+
+// Known statuses per entity, so counts are zero-filled and every status is
+// always present in the metrics (a flat 0 line beats a gap on the dashboard).
+var (
+	taskStatuses   = []string{string(domain.TaskPending), string(domain.TaskLeased), string(domain.TaskRunning), string(domain.TaskCompleted), string(domain.TaskFailed), string(domain.TaskCancelled)}
+	jobStatuses    = []string{string(domain.JobPending), string(domain.JobRunning), string(domain.JobReducing), string(domain.JobCompleted), string(domain.JobFailed), string(domain.JobCancelled)}
+	workerStatuses = []string{string(domain.WorkerOnline), string(domain.WorkerBusy), string(domain.WorkerOffline)}
+)
+
+// StatsRepo answers the aggregate status counts the business metrics report. It
+// runs one cheap GROUP BY per entity; the collector calls this on every scrape.
+type StatsRepo struct {
+	pool *pgxpool.Pool
+}
+
+func NewStatsRepo(pool *pgxpool.Pool) *StatsRepo {
+	return &StatsRepo{pool: pool}
+}
+
+// Counts returns status->count maps for tasks, jobs, and workers, each
+// zero-filled across its known statuses.
+func (r *StatsRepo) Counts(ctx context.Context) (tasks, jobs, workers map[string]int, err error) {
+	if tasks, err = r.countByStatus(ctx, "tasks", taskStatuses); err != nil {
+		return nil, nil, nil, err
+	}
+	if jobs, err = r.countByStatus(ctx, "jobs", jobStatuses); err != nil {
+		return nil, nil, nil, err
+	}
+	if workers, err = r.countByStatus(ctx, "workers", workerStatuses); err != nil {
+		return nil, nil, nil, err
+	}
+	return tasks, jobs, workers, nil
+}
+
+func (r *StatsRepo) countByStatus(ctx context.Context, table string, known []string) (map[string]int, error) {
+	out := make(map[string]int, len(known))
+	for _, s := range known {
+		out[s] = 0 // zero-fill
+	}
+	// table is a fixed internal constant, never user input — safe to format.
+	rows, err := r.pool.Query(ctx, fmt.Sprintf("SELECT status, count(*) FROM %s GROUP BY status", table))
+	if err != nil {
+		return nil, fmt.Errorf("count %s by status: %w", table, err)
+	}
+	defer rows.Close()
+	for rows.Next() {
+		var status string
+		var n int
+		if err := rows.Scan(&status, &n); err != nil {
+			return nil, err
+		}
+		out[status] = n // an unknown status still shows up, which is a useful signal
+	}
+	return out, rows.Err()
+}

coordinator/internal/storage/postgres/task_repo.go

@@ -84,6 +84,9 @@ WITH candidate AS (
     WHERE status = 'pending'
       AND attempt < max_attempts
       AND (cardinality($1::text[]) = 0 OR workload = ANY($1))
+      AND ($5::uuid IS NULL OR NOT EXISTS (
+          SELECT 1 FROM task_results tr
+          WHERE tr.task_id = tasks.id AND tr.owner_id = $5))
     ORDER BY created_at, chunk_index
     FOR UPDATE SKIP LOCKED
     LIMIT 1
@@ -108,7 +111,7 @@ func (r *TaskRepo) ClaimNext(ctx context.Context, f usecase.ClaimFilter) (*domai
 
 	var task *domain.Task
 	err := withRetry(ctx, func(ctx context.Context) error {
-		row := conn(ctx, r.pool).QueryRow(ctx, claimNextSQL, workloads, f.Owner, f.LeaseUntil, f.Now)
+		row := conn(ctx, r.pool).QueryRow(ctx, claimNextSQL, workloads, f.Owner, f.LeaseUntil, f.Now, f.VoterOwner)
 		t, err := scanTask(row)
 		if errors.Is(err, pgx.ErrNoRows) {
 			task = nil
@@ -337,17 +340,26 @@ SET status           = CASE WHEN attempt < max_attempts THEN 'pending'::task_sta
                             ELSE error_message END,
     completed_at     = CASE WHEN attempt >= max_attempts THEN $1 ELSE completed_at END,
     version          = version + 1
-WHERE status IN ('leased','running') AND lease_expires_at < $1`
+WHERE status IN ('leased','running') AND lease_expires_at < $1
+RETURNING job_id`
 
-func (r *TaskRepo) ExpireLeases(ctx context.Context, now time.Time) (int64, error) {
-	var affected int64
+func (r *TaskRepo) ExpireLeases(ctx context.Context, now time.Time) ([]uuid.UUID, error) {
+	var affected []uuid.UUID
 	err := withRetry(ctx, func(ctx context.Context) error {
-		tag, err := conn(ctx, r.pool).Exec(ctx, expireLeasesSQL, now, domain.ErrCodeLeaseExpired)
+		rows, err := conn(ctx, r.pool).Query(ctx, expireLeasesSQL, now, domain.ErrCodeLeaseExpired)
 		if err != nil {
 			return err
 		}
-		affected = tag.RowsAffected()
-		return nil
+		defer rows.Close()
+		affected = affected[:0]
+		for rows.Next() {
+			var jobID uuid.UUID
+			if err := rows.Scan(&jobID); err != nil {
+				return err
+			}
+			affected = append(affected, jobID)
+		}
+		return rows.Err()
 	})
 	return affected, err
 }

coordinator/internal/storage/postgres/task_result_repo.go

@@ -0,0 +1,45 @@
+package postgres
+
+import (
+	"context"
+	"fmt"
+
+	"github.com/google/uuid"
+	"github.com/jackc/pgx/v5/pgxpool"
+)
+
+// TaskResultRepo records and tallies quorum votes for untrusted task results.
+type TaskResultRepo struct {
+	pool *pgxpool.Pool
+}
+
+func NewTaskResultRepo(pool *pgxpool.Pool) *TaskResultRepo {
+	return &TaskResultRepo{pool: pool}
+}
+
+// RecordVote stores (or replaces) one owner's vote for a task's result.
+func (r *TaskResultRepo) RecordVote(ctx context.Context, taskID, ownerID uuid.UUID, sha256 string, artifactID uuid.UUID) error {
+	const sql = `
+INSERT INTO task_results (task_id, owner_id, result_sha256, result_artifact_id)
+VALUES ($1, $2, $3, $4)
+ON CONFLICT (task_id, owner_id) DO UPDATE
+SET result_sha256      = EXCLUDED.result_sha256,
+    result_artifact_id = EXCLUDED.result_artifact_id,
+    created_at         = now()`
+	if _, err := conn(ctx, r.pool).Exec(ctx, sql, taskID, ownerID, sha256, artifactID); err != nil {
+		return fmt.Errorf("record vote: %w", err)
+	}
+	return nil
+}
+
+// CountAgreeing returns how many distinct owners have voted for the given result
+// hash on this task — the size of the agreeing set the quorum is measured
+// against.
+func (r *TaskResultRepo) CountAgreeing(ctx context.Context, taskID uuid.UUID, sha256 string) (int, error) {
+	const sql = `SELECT count(DISTINCT owner_id) FROM task_results WHERE task_id = $1 AND result_sha256 = $2`
+	var n int
+	if err := conn(ctx, r.pool).QueryRow(ctx, sql, taskID, sha256).Scan(&n); err != nil {
+		return 0, fmt.Errorf("count agreeing: %w", err)
+	}
+	return n, nil
+}

coordinator/internal/storage/postgres/ui_read_repo.go

@@ -24,11 +24,15 @@ func (r *UIReadRepo) GetJob(ctx context.Context, id uuid.UUID) (*domain.Job, err
 	return job, err
 }
 
-func (r *UIReadRepo) ListJobs(ctx context.Context, limit int) ([]domain.Job, error) {
+func (r *UIReadRepo) ListJobs(ctx context.Context, owner *uuid.UUID, limit int) ([]domain.Job, error) {
 	if limit < 1 || limit > 100 {
 		return nil, domain.ErrInvalidInput
 	}
-	sql, args, err := psql.Select(jobColumns...).From("jobs").OrderBy("created_at DESC", "id DESC").Limit(uint64(limit)).ToSql()
+	q := psql.Select(jobColumns...).From("jobs")
+	if owner != nil {
+		q = q.Where(sq.Eq{"owner_id": *owner})
+	}
+	sql, args, err := q.OrderBy("created_at DESC", "id DESC").Limit(uint64(limit)).ToSql()
 	if err != nil {
 		return nil, err
 	}
@@ -41,13 +45,13 @@ func (r *UIReadRepo) ListJobs(ctx context.Context, limit int) ([]domain.Job, err
 	for rows.Next() {
 		var j domain.Job
 		var status string
-		var inputURI *string
-		if err := rows.Scan(&j.ID, &j.Workload, &inputURI, &j.Parameters, &status, &j.CreatedAt, &j.CompletedAt); err != nil {
+		if err := rows.Scan(
+			&j.ID, &j.Workload, &j.InputURI, &j.Parameters, &status, &j.CreatedAt, &j.CompletedAt,
+			&j.InputArtifactID, &j.ResultArtifactID, &j.ErrorCode, &j.ErrorMessage, &j.ReducerStartedAt,
+			&j.OwnerID,
+		); err != nil {
 			return nil, err
 		}
-		if inputURI != nil {
-			j.InputURI = *inputURI
-		}
 		j.Status = domain.JobStatus(status)
 		jobs = append(jobs, j)
 	}
@@ -75,6 +79,31 @@ func (r *UIReadRepo) ListTasksByJob(ctx context.Context, jobID uuid.UUID) ([]dom
 	return tasks, rows.Err()
 }
 
+func (r *UIReadRepo) ListTasksByJobs(ctx context.Context, jobIDs []uuid.UUID) (map[uuid.UUID][]domain.Task, error) {
+	out := make(map[uuid.UUID][]domain.Task, len(jobIDs))
+	if len(jobIDs) == 0 {
+		return out, nil
+	}
+	sql, args, err := psql.Select(taskColumns...).From("tasks").
+		Where(sq.Eq{"job_id": jobIDs}).OrderBy("job_id ASC", "chunk_index ASC").ToSql()
+	if err != nil {
+		return nil, err
+	}
+	rows, err := conn(ctx, r.pool).Query(ctx, sql, args...)
+	if err != nil {
+		return nil, fmt.Errorf("list tasks by jobs: %w", err)
+	}
+	defer rows.Close()
+	for rows.Next() {
+		task, err := scanTask(rows)
+		if err != nil {
+			return nil, err
+		}
+		out[task.JobID] = append(out[task.JobID], *task)
+	}
+	return out, rows.Err()
+}
+
 func (r *UIReadRepo) ListWorkers(ctx context.Context, limit int) ([]domain.Worker, error) {
 	if limit < 1 || limit > 100 {
 		return nil, domain.ErrInvalidInput

coordinator/internal/storage/postgres/worker_repo.go

@@ -23,13 +23,13 @@ func NewWorkerRepo(pool *pgxpool.Pool) *WorkerRepo {
 	return &WorkerRepo{pool: pool}
 }
 
-var workerColumns = []string{"id", "name", "capabilities", "status", "last_heartbeat_at", "created_at", "updated_at"}
+var workerColumns = []string{"id", "name", "capabilities", "status", "owner_id", "trust_level", "last_heartbeat_at", "created_at", "updated_at"}
 
 func (r *WorkerRepo) Insert(ctx context.Context, w *domain.Worker) error {
 	sql, args, err := psql.Insert("workers").
 		Columns(workerColumns...).
 		// capabilities is a jsonb column; pgx marshals the []string to a JSON array.
-		Values(w.ID, w.Name, w.Capabilities, string(w.Status),
+		Values(w.ID, w.Name, w.Capabilities, string(w.Status), w.OwnerID, string(w.TrustLevel),
 			w.LastHeartbeatAt, w.CreatedAt, w.UpdatedAt).
 		ToSql()
 	if err != nil {
@@ -95,11 +95,13 @@ func scanWorker(row pgx.Row) (*domain.Worker, error) {
 	var (
 		w      domain.Worker
 		status string
+		trust  string
 	)
-	if err := row.Scan(&w.ID, &w.Name, &w.Capabilities, &status,
+	if err := row.Scan(&w.ID, &w.Name, &w.Capabilities, &status, &w.OwnerID, &trust,
 		&w.LastHeartbeatAt, &w.CreatedAt, &w.UpdatedAt); err != nil {
 		return nil, err
 	}
 	w.Status = domain.WorkerStatus(status)
+	w.TrustLevel = domain.WorkerTrust(trust)
 	return &w, nil
 }

coordinator/internal/token/verifier.go

@@ -0,0 +1,60 @@
+// Package token verifies the HS256 JWTs minted by the userservice. The
+// coordinator only ever *verifies* — it never issues — so this is a deliberately
+// small counterpart to the userservice's issuer. Verification is local: the
+// shared secret is enough, with no runtime call back to the userservice.
+package token
+
+import (
+	"fmt"
+
+	"github.com/golang-jwt/jwt/v5"
+	"github.com/google/uuid"
+)
+
+// Claims is the subset of a userservice token the coordinator cares about.
+type Claims struct {
+	UserID   uuid.UUID
+	Role     string
+	Verified bool
+}
+
+// Verifier checks tokens against the shared HS256 secret.
+type Verifier struct {
+	secret []byte
+}
+
+// NewVerifier returns a Verifier, or nil when secret is empty — a nil Verifier
+// means user-JWT auth is disabled and only the shared service token is accepted.
+func NewVerifier(secret string) *Verifier {
+	if secret == "" {
+		return nil
+	}
+	return &Verifier{secret: []byte(secret)}
+}
+
+type claims struct {
+	Role     string `json:"role"`
+	Verified bool   `json:"verified"`
+	jwt.RegisteredClaims
+}
+
+// Verify checks the signature and expiry and returns the identity. It pins the
+// algorithm to HMAC, rejecting a token that asks for "none" or an RS256 public
+// key — the classic algorithm-substitution attack.
+func (v *Verifier) Verify(raw string) (Claims, error) {
+	var c claims
+	_, err := jwt.ParseWithClaims(raw, &c, func(t *jwt.Token) (any, error) {
+		if _, ok := t.Method.(*jwt.SigningMethodHMAC); !ok {
+			return nil, fmt.Errorf("unexpected signing method: %v", t.Header["alg"])
+		}
+		return v.secret, nil
+	})
+	if err != nil {
+		return Claims{}, err
+	}
+	id, err := uuid.Parse(c.Subject)
+	if err != nil {
+		return Claims{}, fmt.Errorf("token subject is not a uuid: %w", err)
+	}
+	return Claims{UserID: id, Role: c.Role, Verified: c.Verified}, nil
+}

coordinator/internal/token/verifier_test.go

@@ -0,0 +1,98 @@
+package token
+
+import (
+	"testing"
+	"time"
+
+	"github.com/golang-jwt/jwt/v5"
+	"github.com/google/uuid"
+)
+
+const secret = "coordinator-verify-secret-32-bytes!!"
+
+func sign(t *testing.T, method jwt.SigningMethod, key any, sub, role string, exp time.Time) string {
+	t.Helper()
+	return signVerified(t, method, key, sub, role, false, exp)
+}
+
+func signVerified(t *testing.T, method jwt.SigningMethod, key any, sub, role string, verified bool, exp time.Time) string {
+	t.Helper()
+	tok := jwt.NewWithClaims(method, claims{
+		Role:     role,
+		Verified: verified,
+		RegisteredClaims: jwt.RegisteredClaims{
+			Subject:   sub,
+			ExpiresAt: jwt.NewNumericDate(exp),
+		},
+	})
+	raw, err := tok.SignedString(key)
+	if err != nil {
+		t.Fatalf("sign: %v", err)
+	}
+	return raw
+}
+
+func TestVerifyCarriesVerifiedClaim(t *testing.T) {
+	v := NewVerifier(secret)
+	raw := signVerified(t, jwt.SigningMethodHS256, []byte(secret), uuid.New().String(), "user", true, time.Now().Add(time.Hour))
+
+	claims, err := v.Verify(raw)
+	if err != nil {
+		t.Fatalf("verify: %v", err)
+	}
+	if !claims.Verified {
+		t.Error("verified claim not read from token")
+	}
+}
+
+func TestNewVerifierNilWhenNoSecret(t *testing.T) {
+	if NewVerifier("") != nil {
+		t.Error("empty secret must yield a nil verifier (auth disabled)")
+	}
+}
+
+func TestVerifyRoundTrip(t *testing.T) {
+	v := NewVerifier(secret)
+	id := uuid.New()
+	raw := sign(t, jwt.SigningMethodHS256, []byte(secret), id.String(), "admin", time.Now().Add(time.Hour))
+
+	claims, err := v.Verify(raw)
+	if err != nil {
+		t.Fatalf("verify: %v", err)
+	}
+	if claims.UserID != id {
+		t.Errorf("UserID = %v, want %v", claims.UserID, id)
+	}
+	if claims.Role != "admin" {
+		t.Errorf("Role = %q, want admin", claims.Role)
+	}
+}
+
+func TestVerifyRejectsExpired(t *testing.T) {
+	v := NewVerifier(secret)
+	raw := sign(t, jwt.SigningMethodHS256, []byte(secret), uuid.New().String(), "user", time.Now().Add(-time.Minute))
+	if _, err := v.Verify(raw); err == nil {
+		t.Error("expired token accepted")
+	}
+}
+
+func TestVerifyRejectsWrongSecret(t *testing.T) {
+	raw := sign(t, jwt.SigningMethodHS256, []byte(secret), uuid.New().String(), "user", time.Now().Add(time.Hour))
+	if _, err := NewVerifier("another-secret-also-at-least-32-byte").Verify(raw); err == nil {
+		t.Error("token verified under the wrong secret")
+	}
+}
+
+func TestVerifyRejectsNoneAlg(t *testing.T) {
+	raw := sign(t, jwt.SigningMethodNone, jwt.UnsafeAllowNoneSignatureType, uuid.New().String(), "admin", time.Now().Add(time.Hour))
+	if _, err := NewVerifier(secret).Verify(raw); err == nil {
+		t.Error("none-signed token accepted")
+	}
+}
+
+func TestVerifyRejectsNonUUIDSubject(t *testing.T) {
+	raw := sign(t, jwt.SigningMethodHS256, []byte(secret), "not-a-uuid", "user", time.Now().Add(time.Hour))
+	if _, err := NewVerifier(secret).Verify(raw); err == nil {
+		t.Error("non-uuid subject accepted")
+	}
+}

coordinator/internal/transport/http/dto.go

@@ -119,6 +119,8 @@ type jobProgressResponse struct {
 	Done      int       `json:"completed"`
 	Failed    int       `json:"failed"`
 	Cancelled int       `json:"cancelled"`
+	ResultURI string    `json:"result_uri,omitempty"`
+	ErrorCode string    `json:"error_code,omitempty"`
 }
 
 type uploadArtifactResponse struct {
@@ -153,7 +155,7 @@ func toClaimedTaskResponse(c domain.ClaimedTask) claimedTaskResponse {
 }
 
 func toJobProgressResponse(p domain.JobProgress) jobProgressResponse {
-	return jobProgressResponse{
+	out := jobProgressResponse{
 		ID:        p.Job.ID,
 		Status:    string(p.DeriveStatus()),
 		Total:     p.Total,
@@ -163,4 +165,11 @@ func toJobProgressResponse(p domain.JobProgress) jobProgressResponse {
 		Failed:    p.Failed,
 		Cancelled: p.Cancelled,
 	}
+	if p.Job.ResultArtifactID != nil && out.Status == string(domain.JobCompleted) {
+		out.ResultURI = "/jobs/" + p.Job.ID.String() + "/result"
+	}
+	if p.Job.ErrorCode != nil {
+		out.ErrorCode = *p.Job.ErrorCode
+	}
+	return out
 }

coordinator/internal/transport/http/dto_test.go

@@ -0,0 +1,25 @@
+package http
+
+import (
+	"testing"
+
+	"github.com/google/uuid"
+
+	"github.com/emil28092005/SciMesh/coordinator/internal/domain"
+)
+
+func TestJobProgressResponseExposesFinalResultOnlyWhenCompleted(t *testing.T) {
+	id := uuid.New()
+	result := uuid.New()
+	progress := domain.JobProgress{Job: domain.Job{
+		ID: id, Status: domain.JobCompleted, ResultArtifactID: &result,
+	}, Total: 1, Done: 1}
+	if got, want := toJobProgressResponse(progress).ResultURI, "/jobs/"+id.String()+"/result"; got != want {
+		t.Fatalf("result URI = %q, want %q", got, want)
+	}
+
+	progress.Job.Status = domain.JobReducing
+	if got := toJobProgressResponse(progress).ResultURI; got != "" {
+		t.Fatalf("reducing job exposes result URI %q", got)
+	}
+}

coordinator/internal/transport/http/handlers.go

@@ -12,6 +12,7 @@ import (
 
 	"github.com/google/uuid"
 
+	"github.com/emil28092005/SciMesh/coordinator/internal/authctx"
 	"github.com/emil28092005/SciMesh/coordinator/internal/domain"
 	"github.com/emil28092005/SciMesh/coordinator/internal/usecase"
 )
@@ -56,10 +57,24 @@ func (s *Server) handleRegister(w http.ResponseWriter, r *http.Request) {
 		return
 	}
 
-	worker, err := s.uc.RegisterWorker.Execute(ctx, usecase.RegisterWorkerInput{
+	// Resolve the worker's trust tier from how the caller authenticated:
+	//   - shared service token (no requester) -> trusted lab worker
+	//   - verified/admin user JWT             -> trusted volunteer
+	//   - plain user JWT                      -> untrusted (quarantined)
+	in := usecase.RegisterWorkerInput{
 		Name:         req.Name,
 		Capabilities: req.Capabilities,
-	})
+		TrustLevel:   domain.WorkerTrusted,
+	}
+	if requester, ok := authctx.From(ctx); ok {
+		id := requester.UserID
+		in.OwnerID = &id
+		if !requester.IsTrusted() {
+			in.TrustLevel = domain.WorkerUntrusted
+		}
+	}
+
+	worker, err := s.uc.RegisterWorker.Execute(ctx, in)
 	if err != nil {
 		s.writeError(w, r, err)
 		return
@@ -79,6 +94,10 @@ func (s *Server) handleClaim(w http.ResponseWriter, r *http.Request) {
 		s.writeError(w, r, domain.ErrInvalidInput)
 		return
 	}
+	if _, err := uuid.Parse(req.WorkerID); err != nil {
+		s.writeError(w, r, domain.ErrInvalidInput)
+		return
+	}
 
 	claimed, err := s.uc.ClaimTask.Execute(ctx, usecase.ClaimTaskInput{
 		WorkerID:  req.WorkerID,
@@ -146,6 +165,12 @@ func (s *Server) handleResult(w http.ResponseWriter, r *http.Request) {
 		s.writeError(w, r, err)
 		return
 	}
+	if s.uc.ReduceJob != nil {
+		if err := s.uc.ReduceJob.Execute(ctx, task.JobID); err != nil {
+			s.writeError(w, r, err)
+			return
+		}
+	}
 	writeJSON(w, http.StatusOK, taskResponse{ID: task.ID, JobID: task.JobID, Status: string(task.Status)})
 }
 
@@ -395,6 +420,25 @@ func (s *Server) handleGetJob(w http.ResponseWriter, r *http.Request) {
 	writeJSON(w, http.StatusOK, toJobProgressResponse(progress))
 }
 
+func (s *Server) handleGetJobResult(w http.ResponseWriter, r *http.Request) {
+	ctx, cancel := s.reqCtx(r)
+	defer cancel()
+	jobID, ok := s.pathUUID(w, r, "job_id")
+	if !ok {
+		return
+	}
+	art, body, err := s.uc.GetJobResult.Execute(ctx, jobID)
+	if err != nil {
+		s.writeError(w, r, err)
+		return
+	}
+	defer func() { _ = body.Close() }()
+	w.Header().Set("Content-Type", art.ContentType)
+	w.Header().Set("Content-Length", strconv.FormatInt(art.SizeBytes, 10))
+	w.Header().Set("Content-Disposition", fmt.Sprintf("attachment; filename=%q", art.Filename))
+	_, _ = io.Copy(w, body)
+}
+
 // handleCancelJob stops all non-terminal shards for an operator-requested job.
 // It is available to both the bearer API and the separately authenticated UI.
 func (s *Server) handleCancelJob(w http.ResponseWriter, r *http.Request) {

coordinator/internal/transport/http/middleware.go

@@ -9,6 +9,9 @@ import (
 	"net/http"
 	"strings"
 	"time"
+
+	"github.com/emil28092005/SciMesh/coordinator/internal/authctx"
+	tokenpkg "github.com/emil28092005/SciMesh/coordinator/internal/token"
 )
 
 type ctxKey string
@@ -41,25 +44,46 @@ func newRequestID() string {
 
 // withAuth enforces the shared bearer token every worker presents.
 // An empty token disables the check (local development only).
-func withAuth(token string) func(http.Handler) http.Handler {
+// withAuth authenticates a request one of two ways. Workers (and legacy
+// submitters) present the shared service token. When user-JWT auth is enabled
+// (verifier != nil), a submitter may instead present a userservice JWT; on
+// success the requester is stamped into the context so the job use cases can
+// record owner_id and enforce ownership. An empty token with no verifier
+// disables auth entirely (dev only).
+func withAuth(token string, verifier *tokenpkg.Verifier) func(http.Handler) http.Handler {
 	return func(next http.Handler) http.Handler {
 		return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
-			if token == "" {
+			if token == "" && verifier == nil {
 				next.ServeHTTP(w, r)
 				return
 			}
 			presented := strings.TrimPrefix(r.Header.Get("Authorization"), "Bearer ")
-			// Constant-time compare: a byte-by-byte early exit would let an
-			// attacker recover the token by timing responses.
-			if subtle.ConstantTimeCompare([]byte(presented), []byte(token)) != 1 {
+
+			// Shared service token: constant-time compare so a byte-by-byte
+			// early exit cannot leak the token through response timing.
+			if token != "" && subtle.ConstantTimeCompare([]byte(presented), []byte(token)) == 1 {
+				next.ServeHTTP(w, r)
+				return
+			}
+
+			// Otherwise try a user JWT, if that path is configured.
+			if verifier != nil && presented != "" {
+				if claims, err := verifier.Verify(presented); err == nil {
+					ctx := authctx.With(r.Context(), authctx.Requester{
+						UserID:   claims.UserID,
+						Role:     claims.Role,
+						Verified: claims.Verified,
+					})
+					next.ServeHTTP(w, r.WithContext(ctx))
+					return
+				}
+			}
+
 			w.Header().Set("WWW-Authenticate", "Bearer")
 			writeJSON(w, http.StatusUnauthorized, errorResponse{
 				Error:     "unauthorized",
 				RequestID: requestIDFrom(r.Context()),
 			})
-				return
-			}
-			next.ServeHTTP(w, r)
 		})
 	}
 }

coordinator/internal/transport/http/server.go

@@ -7,8 +7,11 @@ import (
 	"context"
 	"log/slog"
 	"net/http"
+	"strings"
 	"time"
 
+	"github.com/emil28092005/SciMesh/coordinator/internal/metrics"
+	tokenpkg "github.com/emil28092005/SciMesh/coordinator/internal/token"
 	"github.com/emil28092005/SciMesh/coordinator/internal/usecase"
 )
 
@@ -22,13 +25,16 @@ type UseCases struct {
 	ClaimTask        *usecase.ClaimTask
 	RenewLease       *usecase.RenewLease
 	CompleteTask     *usecase.CompleteTask
+	ReduceJob        *usecase.ReduceJob
 	FailTask         *usecase.FailTask
 	GetJobStatus     *usecase.GetJobStatus
+	GetJobResult     *usecase.GetJobResult
 	CancelJob        *usecase.CancelJob
 	UploadArtifact   *usecase.UploadArtifact
 	DownloadArtifact *usecase.DownloadArtifact
 	GetTaskInput     *usecase.GetTaskInput
 	Dashboard        *usecase.Dashboard
+	PreviewArtifact  *usecase.PreviewArtifact
 }
 
 type Server struct {
@@ -37,23 +43,47 @@ type Server struct {
 	requestTimeout    time.Duration
 	heartbeatInterval time.Duration
 	maxUploadBytes    int64
+	// verifier validates userservice JWTs. nil disables user-JWT auth, leaving
+	// only the shared service token — the pre-userservice behaviour.
+	verifier *tokenpkg.Verifier
+	// userserviceURL is the base URL the UI proxies login/registration to. Empty
+	// keeps the static basic-auth UI.
+	userserviceURL string
+	// httpClient makes the login/register calls to the userservice.
+	httpClient *http.Client
+	// metrics holds the Prometheus registry and HTTP instrumentation.
+	metrics *metrics.Metrics
 	// ready probes downstream dependencies (the database) for /health. Kept as
 	// a func so the transport layer never imports pgx.
 	ready func(context.Context) error
 }
 
 func NewServer(uc UseCases, log *slog.Logger, requestTimeout, heartbeatInterval time.Duration,
-	maxUploadBytes int64, ready func(context.Context) error) *Server {
+	maxUploadBytes int64, jwtSecret, userserviceURL string, m *metrics.Metrics, ready func(context.Context) error) *Server {
+	if m == nil {
+		m = metrics.New()
+	}
 	return &Server{
 		uc:                uc,
 		log:               log,
 		requestTimeout:    requestTimeout,
 		heartbeatInterval: heartbeatInterval,
 		maxUploadBytes:    maxUploadBytes,
+		verifier:          tokenpkg.NewVerifier(jwtSecret),
+		userserviceURL:    strings.TrimRight(userserviceURL, "/"),
+		httpClient:        &http.Client{Timeout: 10 * time.Second},
+		metrics:           m,
 		ready:             ready,
 	}
 }
 
+// uiSessionMode reports whether the operator UI authenticates via userservice
+// login (cookie session) rather than the static basic-auth token. It needs both
+// a verifier (to check the JWT locally) and a userservice URL (to issue it).
+func (s *Server) uiSessionMode() bool {
+	return s.verifier != nil && s.userserviceURL != ""
+}
+
 // Handler builds the router. Go 1.22's ServeMux matches on method and path
 // wildcards, so no third-party router is needed.
 func (s *Server) Handler(token string, uiToken ...string) http.Handler {
@@ -62,6 +92,7 @@ func (s *Server) Handler(token string, uiToken ...string) http.Handler {
 	protected.HandleFunc("POST /jobs", s.handleCreateJob)
 	protected.HandleFunc("POST /jobs/upload", s.handleUploadDataset)
 	protected.HandleFunc("GET /jobs/{job_id}", s.handleGetJob)
+	protected.HandleFunc("GET /jobs/{job_id}/result", s.handleGetJobResult)
 	protected.HandleFunc("POST /jobs/{job_id}/cancel", s.handleCancelJob)
 	protected.HandleFunc("POST /tasks/claim", s.handleClaim)
 	protected.HandleFunc("GET /tasks/{task_id}/input", s.handleGetTaskInput)
@@ -73,17 +104,57 @@ func (s *Server) Handler(token string, uiToken ...string) http.Handler {
 
 	mux := http.NewServeMux()
 	mux.HandleFunc("GET /health", s.handleHealth)
-	if len(uiToken) > 0 && uiToken[0] != "" && s.uc.Dashboard != nil {
+	// Unauthenticated like /health, so a Prometheus scraper needs no credential.
+	mux.Handle("GET /metrics", s.metrics.Handler())
+
+	hasBasicAuth := len(uiToken) > 0 && uiToken[0] != ""
+	if s.uc.Dashboard != nil && (s.uiSessionMode() || hasBasicAuth) {
 		ui := http.NewServeMux()
-		ui.HandleFunc("GET /ui", s.handleUIHome)
-		ui.HandleFunc("GET /ui/jobs/new", s.handleUINewJob)
-		ui.HandleFunc("GET /ui/jobs/{job_id}", s.handleUIJob)
-		ui.HandleFunc("GET /ui/api/jobs/{job_id}", s.handleUIJobJSON)
-		ui.HandleFunc("POST /ui/api/jobs/{job_id}/cancel", s.handleCancelJob)
-		ui.HandleFunc("POST /ui/api/jobs/upload", s.handleUploadDataset)
-		ui.HandleFunc("GET /ui/jobs/{job_id}/artifacts/{artifact_id}", s.handleUIArtifactDownload)
-		mux.Handle("/ui", chain(ui, withRequestID, withAccessLog(s.log), withBasicAuth(uiToken[0]), withSameOrigin))
-		mux.Handle("/ui/", chain(ui, withRequestID, withAccessLog(s.log), withBasicAuth(uiToken[0]), withSameOrigin))
+
+		// The operator application routes, all requiring an authenticated caller.
+		app := []struct {
+			pattern string
+			handler http.HandlerFunc
+		}{
+			{"GET /ui", s.handleUIHome},
+			{"GET /ui/jobs/new", s.handleUINewJob},
+			{"GET /ui/jobs/{job_id}", s.handleUIJob},
+			{"GET /ui/api/overview", s.handleUIOverviewJSON},
+			{"GET /ui/api/jobs/{job_id}", s.handleUIJobJSON},
+			{"POST /ui/api/jobs/{job_id}/cancel", s.handleCancelJob},
+			{"POST /ui/api/jobs/upload", s.handleUploadDataset},
+			{"GET /ui/jobs/{job_id}/artifacts/{artifact_id}", s.handleUIArtifactDownload},
+			{"GET /ui/jobs/{job_id}/artifacts/{artifact_id}/preview", s.handleUIArtifactPreview},
+		}
+
+		if s.uiSessionMode() {
+			// Public auth pages — reachable without a session so a user can log in.
+			ui.HandleFunc("GET /ui/login", s.handleUILoginForm)
+			ui.HandleFunc("POST /ui/login", s.handleUILogin)
+			ui.HandleFunc("GET /ui/register", s.handleUIRegisterForm)
+			ui.HandleFunc("POST /ui/register", s.handleUIRegister)
+			ui.HandleFunc("POST /ui/logout", s.handleUILogout)
+			gate := withUISession(s.verifier)
+			for _, rt := range app {
+				ui.Handle(rt.pattern, gate(rt.handler))
+			}
+			ui.Handle("GET /ui/profile", gate(http.HandlerFunc(s.handleUIProfile)))
+			// Admin panel: session + admin role.
+			ui.Handle("GET /ui/admin", chain(http.HandlerFunc(s.handleUIAdmin), gate, requireAdmin))
+			ui.Handle("POST /ui/admin/user-action", chain(http.HandlerFunc(s.handleUIAdminUserAction), gate, requireAdmin))
+		} else {
+			for _, rt := range app {
+				ui.HandleFunc(rt.pattern, rt.handler)
+			}
+		}
+
+		common := []func(http.Handler) http.Handler{withRequestID, withAccessLog(s.log)}
+		if !s.uiSessionMode() {
+			common = append(common, withBasicAuth(uiToken[0]))
+		}
+		common = append(common, withSameOrigin)
+		mux.Handle("/ui", chain(ui, common...))
+		mux.Handle("/ui/", chain(ui, common...))
 	} else {
 		// More specific than the protected catch-all: UI absence is not an auth
 		// failure and does not disclose that a UI feature is configured elsewhere.
@@ -93,9 +164,10 @@ func (s *Server) Handler(token string, uiToken ...string) http.Handler {
 	mux.Handle("/", chain(protected,
 		withRequestID,        // outermost: every response gets an ID,
 		withAccessLog(s.log), // including the 401s below
-		withAuth(token),
+		withAuth(token, s.verifier),
 	))
-	return mux
+	// Measure every request once, outermost, with a normalized route label.
+	return s.metrics.Middleware(mux)
 }
 
 // handleHealth reports readiness. It probes the database so an orchestrator

coordinator/internal/transport/http/server_test.go

@@ -25,6 +25,7 @@ const uiToken = "ui-secret"
 type env struct {
 	ts       *httptest.Server
 	blobs    *memstore.BlobStore
+	workerID string
 }
 
 func newEnv(t *testing.T, ready func(context.Context) error) *env {
@@ -41,26 +42,36 @@ func newEnvWithUIToken(t *testing.T, ready func(context.Context) error, configur
 	clk := memstore.NewClock(time.Date(2026, 7, 21, 12, 0, 0, 0, time.UTC))
 	tx := memstore.Tx{}
 	lease := 2 * time.Minute
+	downloadArtifact := usecase.NewDownloadArtifact(arts, blobs)
 
 	uc := coordhttp.UseCases{
 		RegisterWorker:   usecase.NewRegisterWorker(work, clk),
 		CreateJob:        usecase.NewCreateJob(jobs, tasks, tx, clk),
-		SubmitDataset:    usecase.NewSubmitDataset(blobs, arts, jobs, tasks, tx, clk),
-		ClaimTask:        usecase.NewClaimTask(tasks, clk, lease),
+		SubmitDataset:    usecase.NewSubmitDataset(blobs, arts, jobs, tasks, tx, clk, 3),
+		ClaimTask:        usecase.NewClaimTask(tasks, jobs, work, tx, clk, lease),
 		RenewLease:       usecase.NewRenewLease(tasks, work, tx, clk, lease),
-		CompleteTask:     usecase.NewCompleteTask(tasks, jobs, arts, tx, clk),
-		FailTask:         usecase.NewFailTask(tasks, jobs, tx, clk),
+		CompleteTask:     usecase.NewCompleteTask(tasks, jobs, arts, work, memstore.NewTaskResultRepo(), tx, clk, 2),
+		ReduceJob:        usecase.NewReduceJob(jobs, tasks, arts, blobs, tx, clk),
+		FailTask:         usecase.NewFailTask(tasks, jobs, work, tx, clk),
 		GetJobStatus:     usecase.NewGetJobStatus(jobs, tasks),
+		GetJobResult:     usecase.NewGetJobResult(jobs, downloadArtifact),
 		CancelJob:        usecase.NewCancelJob(jobs, tasks, tx, clk),
-		UploadArtifact:   usecase.NewUploadArtifact(tasks, arts, blobs, clk),
-		DownloadArtifact: usecase.NewDownloadArtifact(arts, blobs),
+		UploadArtifact:   usecase.NewUploadArtifact(tasks, work, arts, blobs, tx, clk),
+		DownloadArtifact: downloadArtifact,
 		GetTaskInput:     usecase.NewGetTaskInput(tasks, arts, blobs),
 		Dashboard:        usecase.NewDashboard(memstore.NewUIReadRepo(jobs, tasks, work, arts)),
+		PreviewArtifact:  usecase.NewPreviewArtifact(memstore.NewUIReadRepo(jobs, tasks, work, arts), blobs),
 	}
-	srv := coordhttp.NewServer(uc, slog.New(slog.NewTextHandler(io.Discard, nil)), 5*time.Second, 15*time.Second, 1<<30, ready)
+	worker, err := uc.RegisterWorker.Execute(context.Background(), usecase.RegisterWorkerInput{
+		Name: "test-worker", Capabilities: []string{"w", "similarity-search"},
+	})
+	if err != nil {
+		t.Fatalf("register test worker: %v", err)
+	}
+	srv := coordhttp.NewServer(uc, slog.New(slog.NewTextHandler(io.Discard, nil)), 5*time.Second, 15*time.Second, 1<<30, "", "", nil, ready)
 	ts := httptest.NewServer(srv.Handler(token, configuredUIToken))
 	t.Cleanup(ts.Close)
-	return &env{ts: ts, blobs: blobs}
+	return &env{ts: ts, blobs: blobs, workerID: worker.ID.String()}
 }
 
 func healthy(context.Context) error { return nil }
@@ -68,6 +79,7 @@ func healthy(context.Context) error { return nil }
 // do sends an authenticated JSON request and returns status + decoded body.
 func (e *env) do(t *testing.T, method, path, body string) (int, map[string]any) {
 	t.Helper()
+	body = strings.ReplaceAll(body, `"worker_id":"w1"`, `"worker_id":"`+e.workerID+`"`)
 	req, _ := http.NewRequestWithContext(context.Background(), method, e.ts.URL+path, strings.NewReader(body))
 	req.Header.Set("Authorization", "Bearer "+token)
 	if body != "" {
@@ -141,11 +153,36 @@ func TestUIRequiresDistinctCredentialAndRendersDashboard(t *testing.T) {
 		t.Fatalf("UI status: %d", resp.StatusCode)
 	}
 	body, _ := io.ReadAll(resp.Body)
-	if !strings.Contains(string(body), "SciMesh operator dashboard") {
+	if !strings.Contains(string(body), "SciMesh control room") {
 		t.Errorf("dashboard body missing title")
 	}
 }
 
+func TestUIOverviewReturnsLiveSafeProjection(t *testing.T) {
+	e := newEnv(t, healthy)
+	code, _ := e.do(t, "POST", "/jobs", `{"workload":"w","input_uri":"s3://in","chunks":[{"chunk_index":0,"input_uri":"s3://c","input_sha256":"sha"}]}`)
+	if code != http.StatusCreated {
+		t.Fatalf("create job: %d", code)
+	}
+	req, _ := http.NewRequestWithContext(context.Background(), "GET", e.ts.URL+"/ui/api/overview", nil)
+	req.SetBasicAuth("operator", uiToken)
+	resp, err := http.DefaultClient.Do(req)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer resp.Body.Close()
+	var overview map[string]any
+	if err := json.NewDecoder(resp.Body).Decode(&overview); err != nil {
+		t.Fatal(err)
+	}
+	if resp.StatusCode != http.StatusOK || overview["active_jobs"].(float64) != 1 || overview["online_workers"].(float64) != 1 {
+		t.Fatalf("overview = (%d, %v)", resp.StatusCode, overview)
+	}
+	if _, leaked := overview["worker_auth_token"]; leaked {
+		t.Fatal("overview must not expose authentication configuration")
+	}
+}
+
 func TestUIDisabledReturnsNotFound(t *testing.T) {
 	e := newEnvWithUIToken(t, healthy, "")
 	resp := e.get(t, "/ui")
@@ -330,6 +367,26 @@ func TestRegisterRejectsNoCapabilities(t *testing.T) {
 	}
 }
 
+func TestClaimRequiresRegisteredWorkerAndUsesStoredCapabilities(t *testing.T) {
+	e := newEnv(t, healthy)
+	if code, _ := e.do(t, "POST", "/tasks/claim", `{"worker_id":"not-a-uuid"}`); code != http.StatusBadRequest {
+		t.Fatalf("invalid worker id claim = %d, want 400", code)
+	}
+	if code, _ := e.do(t, "POST", "/tasks/claim", `{"worker_id":"11111111-1111-4111-8111-111111111111"}`); code != http.StatusNotFound {
+		t.Fatalf("unregistered worker claim = %d, want 404", code)
+	}
+	if code, _ := e.do(t, "POST", "/jobs", `{"workload":"w","input_uri":"s3://in","chunks":[{"chunk_index":0,"input_uri":"s3://c","input_sha256":"sha"}]}`); code != http.StatusCreated {
+		t.Fatalf("create job = %d", code)
+	}
+	code, worker := e.do(t, "POST", "/workers/register", `{"name":"search-only","capabilities":["similarity-search"]}`)
+	if code != http.StatusCreated {
+		t.Fatalf("register = %d", code)
+	}
+	if code, _ := e.do(t, "POST", "/tasks/claim", `{"worker_id":"`+worker["worker_id"].(string)+`","capabilities":["w"]}`); code != http.StatusNoContent {
+		t.Fatalf("forged capability claim = %d, want 204", code)
+	}
+}
+
 func TestFullLifecycle(t *testing.T) {
 	e := newEnv(t, healthy)
 
@@ -372,6 +429,113 @@ func TestFullLifecycle(t *testing.T) {
 	}
 }
 
+func TestSimilaritySearchLifecyclePublishesFinalResult(t *testing.T) {
+	e := newEnv(t, healthy)
+	code, job := e.uploadDataset(t, "similarity-search", 10, "chembl_id\tcanonical_smiles\nA\tCC\n")
+	if code != http.StatusCreated {
+		t.Fatalf("upload job: %d", code)
+	}
+	jobID := job["job_id"].(string)
+	code, claim := e.do(t, "POST", "/tasks/claim", `{"worker_id":"w1","capabilities":["similarity-search"]}`)
+	if code != http.StatusOK {
+		t.Fatalf("claim: %d", code)
+	}
+	taskID := claim["task_id"].(string)
+	attempt := int(claim["attempt"].(float64))
+	artifactID := e.putArtifact(t, taskID, "w1", attempt, "rank,chembl_id,canonical_smiles,similarity\n1,A,CC,0.900000\n")
+	if code, _ := e.do(t, "POST", "/tasks/"+taskID+"/result",
+		`{"worker_id":"w1","attempt":`+itoa(attempt)+`,"result":{"artifact_id":"`+artifactID+`"}}`); code != http.StatusOK {
+		t.Fatalf("complete: %d", code)
+	}
+
+	code, progress := e.do(t, "GET", "/jobs/"+jobID, "")
+	if code != http.StatusOK || progress["status"] != "completed" || progress["result_uri"] != "/jobs/"+jobID+"/result" {
+		t.Fatalf("progress = (%d, %v)", code, progress)
+	}
+	req, _ := http.NewRequestWithContext(context.Background(), "GET", e.ts.URL+progress["result_uri"].(string), nil)
+	req.Header.Set("Authorization", "Bearer "+token)
+	resp, err := http.DefaultClient.Do(req)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer resp.Body.Close()
+	body, _ := io.ReadAll(resp.Body)
+	if resp.StatusCode != http.StatusOK || string(body) != "rank,chembl_id,canonical_smiles,similarity\n1,A,CC,0.900000\n" {
+		t.Fatalf("final result = (%d, %q)", resp.StatusCode, body)
+	}
+
+	uiRequest, _ := http.NewRequestWithContext(context.Background(), "GET", e.ts.URL+"/ui/jobs/"+jobID, nil)
+	uiRequest.SetBasicAuth("operator", uiToken)
+	uiResponse, err := http.DefaultClient.Do(uiRequest)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer uiResponse.Body.Close()
+	uiBody, _ := io.ReadAll(uiResponse.Body)
+	if uiResponse.StatusCode != http.StatusOK || !strings.Contains(string(uiBody), "Final result ready") || !strings.Contains(string(uiBody), "Preview CSV") || !strings.Contains(string(uiBody), "Processing speed") {
+		t.Fatalf("final UI = (%d, %q)", uiResponse.StatusCode, uiBody)
+	}
+
+	jsonRequest, _ := http.NewRequestWithContext(context.Background(), "GET", e.ts.URL+"/ui/api/jobs/"+jobID, nil)
+	jsonRequest.SetBasicAuth("operator", uiToken)
+	jsonResponse, err := http.DefaultClient.Do(jsonRequest)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer jsonResponse.Body.Close()
+	var detail map[string]any
+	if err := json.NewDecoder(jsonResponse.Body).Decode(&detail); err != nil {
+		t.Fatal(err)
+	}
+	if jsonResponse.StatusCode != http.StatusOK || detail["final_result_available"] != true {
+		t.Fatalf("final UI JSON = (%d, %v)", jsonResponse.StatusCode, detail)
+	}
+	artifacts := detail["artifacts"].([]any)
+	var finalID string
+	for _, raw := range artifacts {
+		artifact := raw.(map[string]any)
+		if artifact["kind"] == "final_result" && artifact["downloadable"] == true {
+			finalID = artifact["id"].(string)
+			break
+		}
+	}
+	if finalID == "" {
+		t.Fatalf("artifacts = %v, want downloadable final result", artifacts)
+	}
+	var inputID string
+	for _, raw := range artifacts {
+		artifact := raw.(map[string]any)
+		if artifact["kind"] == "input" {
+			inputID = artifact["id"].(string)
+			break
+		}
+	}
+	if inputID == "" {
+		t.Fatalf("artifacts = %v, want input artifact", artifacts)
+	}
+	inputRequest, _ := http.NewRequestWithContext(context.Background(), "GET", e.ts.URL+"/ui/jobs/"+jobID+"/artifacts/"+inputID, nil)
+	inputRequest.SetBasicAuth("operator", uiToken)
+	inputResponse, err := http.DefaultClient.Do(inputRequest)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer inputResponse.Body.Close()
+	if inputResponse.StatusCode != http.StatusNotFound {
+		t.Fatalf("UI input download = %d, want 404", inputResponse.StatusCode)
+	}
+	previewRequest, _ := http.NewRequestWithContext(context.Background(), "GET", e.ts.URL+"/ui/jobs/"+jobID+"/artifacts/"+finalID+"/preview", nil)
+	previewRequest.SetBasicAuth("operator", uiToken)
+	previewResponse, err := http.DefaultClient.Do(previewRequest)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer previewResponse.Body.Close()
+	previewBody, _ := io.ReadAll(previewResponse.Body)
+	if previewResponse.StatusCode != http.StatusOK || !strings.Contains(string(previewBody), "Final result preview") || !strings.Contains(string(previewBody), "0.900000") {
+		t.Fatalf("final preview = (%d, %q)", previewResponse.StatusCode, previewBody)
+	}
+}
+
 func TestForeignArtifactResultConflict(t *testing.T) {
 	e := newEnv(t, healthy)
 	e.do(t, "POST", "/jobs", `{"workload":"w","input_uri":"s3://in",
@@ -393,9 +557,9 @@ func TestForeignArtifactResultConflict(t *testing.T) {
 
 func TestUploadDatasetChunksAndServesInput(t *testing.T) {
 	e := newEnv(t, healthy)
-	tsv := "id\tsmiles\nA\tCC\nB\tCCC\nC\tCCCC\nD\tCCCCC\nE\tCCCCCC\n"
+	tsv := "chembl_id\tcanonical_smiles\nA\tCC\nB\tCCC\nC\tCCCC\nD\tCCCCC\nE\tCCCCCC\n"
 
-	code, body := e.uploadDataset(t, "w", 2, tsv)
+	code, body := e.uploadDataset(t, "similarity-search", 2, tsv)
 	if code != 201 {
 		t.Fatalf("upload: status = %d", code)
 	}
@@ -421,7 +585,7 @@ func TestUploadDatasetChunksAndServesInput(t *testing.T) {
 		t.Fatalf("get input: status = %d", resp.StatusCode)
 	}
 	shard, _ := io.ReadAll(resp.Body)
-	if !strings.HasPrefix(string(shard), "id\tsmiles\n") {
+	if !strings.HasPrefix(string(shard), "chembl_id\tcanonical_smiles\n") {
 		t.Errorf("shard missing header: %q", shard)
 	}
 }
@@ -430,11 +594,12 @@ func TestUploadDatasetLimitsRows(t *testing.T) {
 	e := newEnv(t, healthy)
 	var buf bytes.Buffer
 	mw := multipart.NewWriter(&buf)
-	_ = mw.WriteField("workload", "w")
+	_ = mw.WriteField("workload", "similarity-search")
+	_ = mw.WriteField("parameters", `{"query_smiles":"CCO"}`)
 	_ = mw.WriteField("chunk_rows", "2")
 	_ = mw.WriteField("max_rows", "3")
 	fw, _ := mw.CreateFormFile("file", "chembl.tsv")
-	_, _ = io.Copy(fw, strings.NewReader("id\tsmiles\nA\tCC\nB\tCCC\nC\tCCCC\nD\tCCCCC\n"))
+	_, _ = io.Copy(fw, strings.NewReader("chembl_id\tcanonical_smiles\nA\tCC\nB\tCCC\nC\tCCCC\nD\tCCCCC\n"))
 	_ = mw.Close()
 	req, _ := http.NewRequestWithContext(context.Background(), "POST", e.ts.URL+"/jobs/upload", &buf)
 	req.Header.Set("Authorization", "Bearer "+token)
@@ -451,6 +616,29 @@ func TestUploadDatasetLimitsRows(t *testing.T) {
 	}
 }
 
+func TestUploadDatasetRejectsMissingChEMBLColumns(t *testing.T) {
+	e := newEnv(t, healthy)
+	var buf bytes.Buffer
+	mw := multipart.NewWriter(&buf)
+	_ = mw.WriteField("workload", "similarity-search")
+	_ = mw.WriteField("parameters", `{"query_smiles":"CCO"}`)
+	_ = mw.WriteField("chunk_rows", "2")
+	fw, _ := mw.CreateFormFile("file", "not-chembl.tsv")
+	_, _ = io.Copy(fw, strings.NewReader("id\tsmiles\nA\tCC\n"))
+	_ = mw.Close()
+	req, _ := http.NewRequestWithContext(context.Background(), "POST", e.ts.URL+"/jobs/upload", &buf)
+	req.Header.Set("Authorization", "Bearer "+token)
+	req.Header.Set("Content-Type", mw.FormDataContentType())
+	resp, err := http.DefaultClient.Do(req)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer resp.Body.Close()
+	if resp.StatusCode != http.StatusBadRequest {
+		t.Errorf("missing ChEMBL columns = %d, want 400", resp.StatusCode)
+	}
+}
+
 func TestErrorMappings(t *testing.T) {
 	e := newEnv(t, healthy)
 	zero := "00000000-0000-0000-0000-000000000000"
@@ -478,10 +666,11 @@ func TestUploadDatasetRejectsAmbiguousMultipartInput(t *testing.T) {
 	e := newEnv(t, healthy)
 	var buf bytes.Buffer
 	mw := multipart.NewWriter(&buf)
-	_ = mw.WriteField("workload", "w")
+	_ = mw.WriteField("workload", "similarity-search")
+	_ = mw.WriteField("parameters", `{"query_smiles":"CCO"}`)
 	_ = mw.WriteField("chunk_rows", "not-a-number")
 	fw, _ := mw.CreateFormFile("file", "chembl.tsv")
-	_, _ = io.Copy(fw, strings.NewReader("id\tsmiles\nA\tCC\n"))
+	_, _ = io.Copy(fw, strings.NewReader("chembl_id\tcanonical_smiles\nA\tCC\n"))
 	_ = mw.Close()
 
 	req, _ := http.NewRequestWithContext(context.Background(), "POST", e.ts.URL+"/jobs/upload", &buf)
@@ -505,6 +694,9 @@ func (e *env) putArtifact(t *testing.T, taskID, worker string, attempt int, data
 		e.ts.URL+"/tasks/"+taskID+"/artifacts/r.csv", strings.NewReader(data))
 	req.Header.Set("Authorization", "Bearer "+token)
 	req.Header.Set("Content-Type", "text/csv")
+	if worker == "w1" {
+		worker = e.workerID
+	}
 	req.Header.Set("X-Worker-ID", worker)
 	req.Header.Set("X-Task-Attempt", itoa(attempt))
 	resp, err := http.DefaultClient.Do(req)
@@ -526,6 +718,7 @@ func (e *env) uploadDataset(t *testing.T, workload string, rows int, tsv string)
 	var buf bytes.Buffer
 	mw := multipart.NewWriter(&buf)
 	_ = mw.WriteField("workload", workload)
+	_ = mw.WriteField("parameters", `{"query_smiles":"CCO"}`)
 	_ = mw.WriteField("chunk_rows", itoa(rows))
 	fw, _ := mw.CreateFormFile("file", "chembl.tsv")
 	_, _ = io.Copy(fw, strings.NewReader(tsv))

coordinator/internal/transport/http/templates/admin.html

@@ -0,0 +1,46 @@
+{{define "admin.html"}}
+<!doctype html>
+<html lang="en">
+<head>
+  <meta charset="utf-8">
+  <meta name="viewport" content="width=device-width, initial-scale=1">
+  <title>Admin · SciMesh</title>
+  <style>
+    :root{color:#e5efff;background:#08111f;font:16px/1.5 Inter,ui-sans-serif,system-ui,sans-serif;color-scheme:dark}*{box-sizing:border-box}body{margin:0;background:radial-gradient(circle at 10% -8%,#183f77 0,transparent 32rem),#08111f}.page{max-width:820px;margin:auto;padding:28px 22px 64px}a{color:#94bdff}.top{display:flex;justify-content:space-between;align-items:center;gap:12px}.eyebrow{margin:0;color:#7baaff;font-size:.78rem;font-weight:800;letter-spacing:.14em;text-transform:uppercase}h1{margin:6px 0 0;color:#f4f8ff;font-size:clamp(1.8rem,4vw,2.6rem);letter-spacing:-.04em}.lead{max-width:640px;margin:10px 0 0;color:#aabed9}.card{margin-top:24px;border:1px solid #294662;border-radius:15px;background:#0d1a2cdd;box-shadow:0 20px 45px #00000021;padding:22px}.card h2{margin:0 0 4px;color:#f1f6ff;font-size:1.1rem}.card p{margin:0;color:#9fb3cf;font-size:.92rem}label{display:block;margin:16px 0 5px;color:#eaf2ff;font-weight:750}input{width:100%;border:1px solid #42617f;border-radius:9px;padding:10px 11px;background:#0a1626;color:#e5efff;font:inherit}input:focus{outline:2px solid #5d97f5;outline-offset:1px}.actions{display:flex;flex-wrap:wrap;gap:9px;margin-top:16px}.btn{border:0;border-radius:9px;padding:10px 14px;font:inherit;font-weight:800;cursor:pointer}.btn-primary{background:#67e3b8;color:#062018}.btn-muted{background:#23344d;color:#dce8ff}.notice{margin-top:16px;border-radius:10px;padding:11px 13px;font-weight:700}.ok{background:#123f34;color:#76efb5}.err{background:#552334;color:#ff9bad}.muted{color:#8ba2c2}.hint{margin-top:4px;color:#92a9c6;font-size:.85rem}</style>
+</head>
+<body>
+  <main class="page">
+    <header class="top">
+      <div><p class="eyebrow">Admin panel</p><h1>User &amp; run control</h1></div>
+      <div style="display:flex;gap:10px;align-items:center"><a href="/ui">← Dashboard</a><a href="/ui/profile">Profile</a><form method="post" action="/ui/logout" style="margin:0"><button class="btn btn-muted" type="submit">Log out</button></form></div>
+    </header>
+    <p class="lead">Signed in as <strong>{{.Role}}</strong>. Promote or verify a user by their id, and control every job from the dashboard.</p>
+
+    {{if .Msg}}<div class="notice ok">{{.Msg}}</div>{{end}}
+    {{if .Error}}<div class="notice err">{{.Error}}</div>{{end}}
+
+    <section class="card">
+      <h2>Manage a user</h2>
+      <p>Paste the user id (the JWT <code>sub</code> / the value shown at registration). Actions are applied immediately.</p>
+      <form method="post" action="/ui/admin/user-action">
+        <label for="user_id">User id</label>
+        <input id="user_id" name="user_id" placeholder="00000000-0000-0000-0000-000000000000" autocomplete="off" required>
+        <p class="hint">Promote makes them an admin; Verify marks them a trusted contributor (their workers skip quorum).</p>
+        <div class="actions">
+          <button class="btn btn-primary" name="action" value="promote" type="submit">Make admin</button>
+          <button class="btn btn-muted" name="action" value="demote" type="submit">Remove admin</button>
+          <button class="btn btn-primary" name="action" value="verify" type="submit">Verify</button>
+          <button class="btn btn-muted" name="action" value="unverify" type="submit">Unverify</button>
+        </div>
+      </form>
+    </section>
+
+    <section class="card">
+      <h2>Jobs &amp; tasks</h2>
+      <p>As an admin you already see <strong>every user's jobs</strong> on the dashboard, with per-task status and job cancellation. A regular user sees only their own.</p>
+      <div class="actions"><a class="btn btn-muted" href="/ui" style="text-decoration:none">Open the dashboard →</a></div>
+    </section>
+  </main>
+</body>
+</html>
+{{end}}

coordinator/internal/transport/http/templates/artifact-preview.html

@@ -0,0 +1,35 @@
+{{define "artifact-preview.html"}}
+<!doctype html>
+<html lang="en">
+<head>
+  <meta charset="utf-8">
+  <meta name="viewport" content="width=device-width, initial-scale=1">
+  <title>SciMesh · artifact preview</title>
+  <style>
+    :root{color:#e4eeff;background:#08111f;font:16px/1.5 Inter,ui-sans-serif,system-ui,sans-serif;color-scheme:dark}body{margin:0;background:radial-gradient(circle at 10% -5%,#173f76 0,transparent 34rem),#08111f}.page{max-width:1180px;margin:auto;padding:32px 20px 56px}a{color:#8ab5ff}.back{text-decoration:none}h1{margin:18px 0 4px;color:#f4f8ff;font-size:1.6rem;word-break:break-word}.muted{color:#9cb0cb}.notice{margin:16px 0;padding:15px 17px;border:1px solid #aa8844;border-radius:10px;background:#302610;color:#f2dd9a}.table-wrap{overflow-x:auto;border:1px solid #294662;border-radius:10px;background:#0d1a2cdc;margin-top:16px}table{width:100%;border-collapse:collapse}td,th{padding:10px 12px;border-bottom:1px solid #203a55;text-align:left;vertical-align:top;white-space:pre-wrap;word-break:break-word}th{color:#9cb9dc;font-size:.78rem;text-transform:uppercase;letter-spacing:.06em;background:#10253d}tr:last-child td{border:0}.empty{padding:24px;text-align:center;color:#9cb0cb}
+  </style>
+</head>
+<body>
+  <main class="page">
+    <a class="back" href="/ui/jobs/{{.JobID}}">← Back to job</a>
+    <h1>Preview: {{.Filename}}</h1>
+    {{if .Diagnostic}}
+      <p class="muted">Diagnostic preview — a shard-level partial result, not the final molecular-search answer. At most {{.RowLimit}} rows and {{.ByteLimit}} bytes are read from storage.</p>
+    {{else}}
+      <p class="muted">Final result preview. At most {{.RowLimit}} rows and {{.ByteLimit}} bytes are read from storage.</p>
+    {{end}}
+    {{if not .Previewable}}
+      <div class="notice">{{.Reason}}</div>
+    {{else}}
+      {{if .Truncated}}<div class="notice">Truncated: showing at most the first {{.RowLimit}} rows or {{.ByteLimit}} bytes. Download the artifact for its full contents.</div>{{end}}
+      <div class="table-wrap">
+        <table>
+          <tr>{{range .Headers}}<th>{{.}}</th>{{end}}</tr>
+          {{range .Rows}}<tr>{{range .}}<td>{{.}}</td>{{end}}</tr>{{else}}<tr><td class="empty" colspan="99">No data rows.</td></tr>{{end}}
+        </table>
+      </div>
+    {{end}}
+  </main>
+</body>
+</html>
+{{end}}

coordinator/internal/transport/http/templates/dashboard.html

@@ -4,20 +4,36 @@
 <head>
   <meta charset="utf-8">
   <meta name="viewport" content="width=device-width, initial-scale=1">
-  <title>SciMesh operator dashboard</title>
+  <title>SciMesh control room</title>
   <style>
-    :root{color:#172033;background:#f6f8fc;font:16px/1.5 system-ui,sans-serif}body{margin:0}.page{max-width:1180px;margin:auto;padding:32px 20px 56px}.top{display:flex;justify-content:space-between;gap:24px;align-items:start}.eyebrow{margin:0;color:#50617d;font-size:.86rem;font-weight:700;text-transform:uppercase;letter-spacing:.08em}h1{margin:.2rem 0;font-size:2rem}h2{margin:32px 0 12px;font-size:1.28rem}.lead{margin:0;color:#56657c}.button{display:inline-block;border:0;border-radius:8px;padding:11px 15px;background:#1f5eff;color:#fff;font-weight:700;text-decoration:none;white-space:nowrap}.notice{margin-top:24px;padding:16px 18px;border:1px solid #f2cb72;border-radius:10px;background:#fff8e6}.notice strong{display:block}.steps{display:grid;grid-template-columns:repeat(3,1fr);gap:12px;margin-top:14px}.step,.card{padding:16px;border:1px solid #dfe5f0;border-radius:10px;background:#fff}.step b{display:block;color:#1f5eff}.table-wrap{overflow-x:auto;background:#fff;border:1px solid #dfe5f0;border-radius:10px}table{width:100%;border-collapse:collapse}td,th{padding:13px 14px;border-bottom:1px solid #e8ecf4;text-align:left;vertical-align:top}th{color:#50617d;font-size:.78rem;text-transform:uppercase;letter-spacing:.06em}tr:last-child td{border:0}a{color:#174ecf}small,.muted{color:#68758b}.status{display:inline-block;border-radius:999px;padding:3px 9px;font-size:.84rem;font-weight:700}.status-success{background:#dff6e9;color:#126b3d}.status-danger{background:#ffe4e6;color:#a31135}.status-active{background:#e4edff;color:#174ecf}.status-waiting{background:#edf0f5;color:#50617d}.bar{height:7px;min-width:120px;margin-top:7px;overflow:hidden;border-radius:999px;background:#e6eaf1}.bar>span{display:block;height:100%;background:#1f5eff}.kicker{font-variant-numeric:tabular-nums}.empty{padding:28px;text-align:center;color:#68758b}.worker{display:grid;grid-template-columns:1.3fr .8fr 2fr 1fr;gap:12px;align-items:center}.worker+.worker{border-top:1px solid #e8ecf4;padding-top:12px;margin-top:12px}@media(max-width:760px){.top,.steps{display:block}.button{margin-top:12px}.step{margin-top:10px}.worker{grid-template-columns:1fr}.hide-mobile{display:none}}
+    :root{color:#dce8ff;background:#08111f;font:16px/1.5 Inter,ui-sans-serif,system-ui,sans-serif;color-scheme:dark}*{box-sizing:border-box}body{margin:0;background:radial-gradient(circle at 15% -10%,#163d77 0,transparent 32rem),radial-gradient(circle at 95% 5%,#123e39 0,transparent 29rem),#08111f}.page{max-width:1240px;margin:auto;padding:28px 22px 64px}.top{display:flex;align-items:flex-start;justify-content:space-between;gap:22px}.eyebrow{margin:0;color:#78a9ff;font-size:.77rem;font-weight:750;letter-spacing:.14em;text-transform:uppercase}.title{max-width:700px;margin:8px 0;font-size:clamp(2rem,5vw,3.6rem);line-height:1.04;letter-spacing:-.055em}.lead{max-width:690px;margin:0;color:#aabbd5;font-size:1.05rem}.button{display:inline-flex;align-items:center;gap:8px;border:0;border-radius:10px;padding:12px 16px;background:#4f8cff;color:#071224;font:inherit;font-weight:800;text-decoration:none;box-shadow:0 12px 30px #163d7766}.live{display:inline-flex;align-items:center;gap:7px;margin-top:18px;color:#8ba2c2;font-size:.87rem}.pulse{width:8px;height:8px;border-radius:50%;background:#5ee6a6;box-shadow:0 0 0 5px #5ee6a622}.summary{display:grid;grid-template-columns:1.4fr repeat(3,1fr);gap:13px;margin:32px 0}.panel,.metric,.flow-step,.job,.worker{border:1px solid #26415f;border-radius:15px;background:#0d1a2cdd;box-shadow:0 20px 45px #0000001f}.panel{padding:20px}.pipeline{display:grid;grid-template-columns:repeat(4,1fr);gap:10px;margin-top:16px}.flow-step{position:relative;padding:14px;min-height:116px}.flow-step:not(:last-child):after{content:"";position:absolute;top:35px;right:-10px;width:10px;height:2px;background:#3c5d83}.flow-step b{display:block;color:#f2f7ff}.flow-step small{display:block;margin-top:6px;color:#91a8c6}.flow-step .dot{display:inline-block;width:9px;height:9px;margin-right:7px;border-radius:50%;background:#5ee6a6}.metric{padding:16px}.metric b{display:block;margin-top:7px;color:#f5f8ff;font-size:2rem;line-height:1}.metric span{color:#9bb0cc;font-size:.84rem}.section-head{display:flex;align-items:baseline;justify-content:space-between;gap:12px;margin:36px 0 12px}.section-head h2{margin:0;color:#f3f7ff;font-size:1.18rem}.section-head p{margin:0;color:#8fa6c3;font-size:.9rem}.jobs{display:grid;gap:10px}.job{display:grid;grid-template-columns:minmax(210px,1.35fr) minmax(150px,.9fr) minmax(180px,1fr) auto;gap:18px;align-items:center;padding:17px 18px;text-decoration:none;color:inherit;transition:border-color .15s,transform .15s}.job:hover{border-color:#5d96ee;transform:translateY(-1px)}.job-name{color:#f3f7ff;font-weight:750}.job-id{margin-top:3px;color:#8196b3;font-family:ui-monospace,SFMono-Regular,monospace;font-size:.76rem}.badge{display:inline-flex;align-items:center;border-radius:999px;padding:4px 9px;font-size:.78rem;font-weight:800}.badge-waiting{background:#23344d;color:#b9cce9}.badge-active{background:#173d77;color:#9fc7ff}.badge-success{background:#123f34;color:#76efb5}.badge-danger{background:#552334;color:#ff9bad}.bar{height:7px;margin-top:8px;overflow:hidden;border-radius:999px;background:#20344e}.bar span{display:block;height:100%;border-radius:inherit;background:linear-gradient(90deg,#5a92ff,#61e1bd)}.numbers{color:#afc0d9;font-size:.88rem}.arrow{color:#79aaff;font-size:1.35rem}.workers{display:grid;grid-template-columns:repeat(3,1fr);gap:11px}.worker{padding:15px}.worker-head{display:flex;justify-content:space-between;gap:8px}.worker strong{color:#f3f7ff}.worker small,.worker p{color:#95a9c4}.worker p{margin:12px 0 0}.cap{display:inline-block;margin:4px 5px 0 0;border:1px solid #365576;border-radius:5px;padding:2px 6px;color:#a9c9f4;font:.75rem ui-monospace,SFMono-Regular,monospace}.empty{padding:30px;border:1px dashed #35516f;border-radius:14px;color:#9ab0cb;text-align:center}.offline{color:#faafbd}.sr{position:absolute;width:1px;height:1px;clip:rect(0,0,0,0);overflow:hidden;white-space:nowrap}@media(max-width:820px){.top,.section-head{display:block}.button{margin-top:18px}.summary{grid-template-columns:1fr 1fr}.panel{grid-column:span 2}.pipeline{grid-template-columns:1fr 1fr}.flow-step:not(:last-child):after{display:none}.job{grid-template-columns:1fr 1fr}.arrow{display:none}.workers{grid-template-columns:1fr 1fr}}@media(max-width:540px){.page{padding:22px 14px}.summary,.workers{grid-template-columns:1fr}.panel{grid-column:auto}.pipeline,.job{grid-template-columns:1fr}.title{font-size:2.35rem}}
   </style>
 </head>
 <body>
   <main class="page">
-    <header class="top"><div><p class="eyebrow">Local coordinator</p><h1>SciMesh operator dashboard</h1><p class="lead">See where a computation is and what should happen next.</p></div><a class="button" href="/ui/jobs/new">Start a check</a></header>
-    <section class="notice" aria-label="Current pipeline limitation"><strong>This screen currently diagnoses shard jobs.</strong><span>Workers upload partial CSVs to the coordinator. Until a reducer is implemented, those files are not one final scientific result.</span><div class="steps"><div class="step"><b>1. Upload TSV</b>The coordinator splits the file into shard tasks.</div><div class="step"><b>2. Wait for a worker</b>A worker claims a shard, calculates similarity, and returns a CSV.</div><div class="step"><b>3. Inspect artifacts</b>Download a partial result from the job page.</div></div></section>
-    <h2>Recent jobs</h2>
-    <div class="table-wrap"><table><tr><th>Computation</th><th>State</th><th>Progress</th><th class="hide-mobile">Created</th></tr>{{range .Jobs}}<tr><td><a href="/ui/jobs/{{.ID}}"><strong>{{workloadLabel .Workload}}</strong></a><br><small>Open job details</small></td><td><span class="status status-{{statusClass .Status}}">{{statusLabel .Status}}</span><br><small>{{statusHint .Status}}</small></td><td class="kicker"><strong>{{.Completed}} / {{.Total}}</strong> complete{{if gt .Failed 0}} · <span style="color:#a31135">failed: {{.Failed}}</span>{{end}}{{if gt .Cancelled 0}} · <span>stopped: {{.Cancelled}}</span>{{end}}<div class="bar"><span style="width:{{progressPercent .Completed .Failed .Cancelled .Total}}%"></span></div></td><td class="hide-mobile"><small>{{time .CreatedAt}}</small></td></tr>{{else}}<tr><td colspan="4" class="empty"><strong>No jobs yet.</strong><br>Click “Start a check”, upload a small TSV, and leave a worker running.</td></tr>{{end}}</table></div>
-    <h2>Workers</h2>
-    <section class="card">{{range .Workers}}<div class="worker"><div><strong>{{.Name}}</strong><br><small>{{.ID}}</small></div><div><span class="status status-{{if eq .Status "online"}}success{{else}}waiting{{end}}">{{workerStatusLabel .Status}}</span></div><div>{{range .Capabilities}}<code>{{.}}</code> {{end}}</div><div class="muted">Last signal<br>{{time .LastHeartbeatAt}}</div></div>{{else}}<div class="empty"><strong>No worker is registered yet.</strong><br>Run <code>scimesh-worker</code> with the coordinator URL and worker token.</div>{{end}}</section>
+    <header class="top">
+      <div><p class="eyebrow">Local scientific compute</p><h1 class="title">SciMesh control room</h1><p class="lead">Follow the real path from a molecular TSV to a globally reduced similarity result—without reading coordinator logs.</p><div class="live"><i class="pulse"></i><span id="refresh-state">Live overview · refreshes every 2 seconds</span></div></div>
+      <div style="display:flex;gap:10px;align-items:center;flex-wrap:wrap">{{if .Session}}<span class="live" style="margin-top:0">Signed in · {{.Session.Role}}</span>{{end}}{{if .Session}}<a class="button" href="/ui/profile" style="background:#23344d;color:#dce8ff;box-shadow:none">Profile</a>{{end}}{{if and .Session (eq .Session.Role "admin")}}<a class="button" href="/ui/admin" style="background:#23344d;color:#dce8ff;box-shadow:none">Admin</a>{{end}}<a class="button" href="/ui/jobs/new">＋ New similarity search</a>{{if .Session}}<form method="post" action="/ui/logout" style="margin:0"><button class="button" type="submit" style="background:#23344d;color:#dce8ff;box-shadow:none">Log out</button></form>{{end}}</div>
+    </header>
+    <section class="summary" aria-label="Pipeline summary">
+      <div class="panel"><strong>How a search becomes a result</strong><div class="pipeline"><div class="flow-step"><span><i class="dot"></i>01</span><b>Upload TSV</b><small>The coordinator validates and slices the dataset.</small></div><div class="flow-step"><span><i class="dot"></i>02</span><b>Run shards</b><small>Workers fingerprint molecules and return shard top-k CSVs.</small></div><div class="flow-step"><span><i class="dot"></i>03</span><b>Merge exactly</b><small>The coordinator ranks retained candidates deterministically.</small></div><div class="flow-step"><span><i class="dot"></i>04</span><b>Download CSV</b><small>A checksum-protected global result is ready.</small></div></div></div>
+      <div class="metric"><span>Active runs</span><b id="active-jobs">{{.ActiveJobs}}</b><small>waiting, running, or merging</small></div>
+      <div class="metric"><span>Available workers</span><b id="online-workers">{{.OnlineWorkers}}</b><small>recently registered</small></div>
+      <div class="metric"><span>Finished runs</span><b id="finished-jobs">{{.FinishedJobs}}</b><small>in the latest 20</small></div>
+    </section>
+
+    <section><div class="section-head"><h2>Recent computations</h2><p id="job-count">{{len .Jobs}} shown · newest first</p></div><div id="jobs" class="jobs">{{range .Jobs}}<a class="job" href="/ui/jobs/{{.ID}}"><div><div class="job-name">{{workloadLabel .Workload}}</div><div class="job-id">{{.ID}}</div></div><div><span class="badge badge-{{statusClass .Status}}">{{statusLabel .Status}}</span><div class="job-id">{{statusHint .Status}}</div></div><div><div class="numbers"><b>{{.Completed}}</b> / {{.Total}} shards complete{{if gt .Failed 0}} · <span class="offline">{{.Failed}} failed</span>{{end}}</div><div class="bar"><span style="width:{{progressPercent .Completed .Failed .Cancelled .Total}}%"></span></div></div><span class="arrow" aria-hidden="true">→</span></a>{{else}}<div class="empty"><strong>No computations yet.</strong><br>Start a small similarity search, then keep one or more workers running to watch this dashboard come alive.</div>{{end}}</div></section>
+    <section><div class="section-head"><h2>Worker fleet</h2><p>Workers register themselves; this page never controls their processes.</p></div><div id="workers" class="workers">{{range .Workers}}<article class="worker"><div class="worker-head"><strong>{{.Name}}</strong><span class="badge badge-{{workerStatusClass .Status}}">{{workerStatusLabel .Status}}</span></div><small>{{.ID}}</small><p>{{range .Capabilities}}<span class="cap">{{.}}</span>{{end}}</p><p>Last signal · {{time .LastHeartbeatAt}}</p></article>{{else}}<div class="empty"><strong>No worker is registered.</strong><br>Start <code>scimesh-worker</code> in another terminal, then return here.</div>{{end}}</div></section>
   </main>
+  <script>
+    const statusInfo={pending:['Waiting for a worker','waiting'],leased:['Assigned to a worker','active'],running:['Running','active'],reducing:['Merging results','active'],completed:['Completed','success'],failed:['Needs attention','danger'],cancelled:['Stopped','waiting']};
+    const pct=j=>j.total?Math.min(100,Math.floor((j.completed+j.failed+j.cancelled)*100/j.total)):0;
+    const node=(tag,text,cls)=>{const n=document.createElement(tag);if(text!==undefined)n.textContent=text;if(cls)n.className=cls;return n};
+    const renderJobs=jobs=>{const box=document.querySelector('#jobs');box.replaceChildren();if(!jobs.length){const empty=node('div',undefined,'empty');empty.append(node('strong','No computations yet.'),document.createElement('br'),document.createTextNode('Start a small similarity search, then keep one or more workers running to watch this dashboard come alive.'));box.append(empty);return}for(const job of jobs){const info=statusInfo[job.status]||[job.status,'waiting'],link=node('a',undefined,'job');link.href='/ui/jobs/'+encodeURIComponent(job.id);const intro=node('div');intro.append(node('div',job.workload==='similarity-search'?'Molecule similarity search':job.workload,'job-name'),node('div',job.id,'job-id'));const state=node('div');state.append(node('span',info[0],'badge badge-'+info[1]),node('div',job.status==='reducing'?'Every shard is complete; coordinator is ranking the global top-k.':'Live coordinator state','job-id'));const progress=node('div'),numbers=node('div',undefined,'numbers');numbers.append(node('b',String(job.completed)),document.createTextNode(' / '+job.total+' shards complete'));if(job.failed){numbers.append(document.createTextNode(' · '),node('span',job.failed+' failed','offline'))}const bar=node('div',undefined,'bar'),fill=node('span');fill.style.width=pct(job)+'%';bar.append(fill);progress.append(numbers,bar);link.append(intro,state,progress,node('span','→','arrow'));box.append(link)}};
+    const renderWorkers=workers=>{const box=document.querySelector('#workers');box.replaceChildren();if(!workers.length){const empty=node('div',undefined,'empty');empty.append(node('strong','No worker is registered.'),document.createElement('br'),document.createTextNode('Start scimesh-worker in another terminal, then return here.'));box.append(empty);return}for(const worker of workers){const card=node('article',undefined,'worker'),head=node('div',undefined,'worker-head'),left=node('div'),workerInfo=worker.status==='online'?['Available','success']:worker.status==='busy'?['Busy','active']:['Offline','waiting'];left.append(node('strong',worker.name),node('small',worker.id));head.append(left,node('span',workerInfo[0],'badge badge-'+workerInfo[1]));const caps=node('p');for(const capability of worker.capabilities||[])caps.append(node('span',capability,'cap'));card.append(head,caps,node('p','Last signal · '+new Date(worker.last_heartbeat_at).toLocaleString()));box.append(card)}};
+    let timer;const refresh=async()=>{try{const response=await fetch('/ui/api/overview',{headers:{Accept:'application/json'}});if(!response.ok)throw Error();const view=await response.json();document.querySelector('#active-jobs').textContent=view.active_jobs;document.querySelector('#online-workers').textContent=view.online_workers;document.querySelector('#finished-jobs').textContent=view.finished_jobs;document.querySelector('#job-count').textContent=view.jobs.length+' shown · newest first';renderJobs(view.jobs);renderWorkers(view.workers);document.querySelector('#refresh-state').textContent='Live overview · updated just now'}catch(_){document.querySelector('#refresh-state').textContent='Connection interrupted · trying again automatically'}};
+    const start=()=>{if(!timer&&!document.hidden)timer=setInterval(refresh,2000)};document.addEventListener('visibilitychange',()=>{if(document.hidden&&timer){clearInterval(timer);timer=undefined}else start()});start();
+  </script>
 </body>
 </html>
 {{end}}

coordinator/internal/transport/http/templates/login.html

@@ -0,0 +1,27 @@
+{{define "login.html"}}
+<!doctype html>
+<html lang="en">
+<head>
+  <meta charset="utf-8">
+  <meta name="viewport" content="width=device-width, initial-scale=1">
+  <title>Sign in · SciMesh</title>
+  <style>
+    :root{color:#e5efff;background:#08111f;font:16px/1.5 Inter,ui-sans-serif,system-ui,sans-serif;color-scheme:dark}*{box-sizing:border-box}body{margin:0;min-height:100vh;display:grid;place-items:center;background:radial-gradient(circle at 10% -8%,#183f77 0,transparent 32rem),#08111f}a{color:#94bdff}.card{width:min(92vw,380px);border:1px solid #294662;border-radius:15px;background:#0d1a2cdd;box-shadow:0 20px 45px #00000021;padding:28px}.eyebrow{margin:0 0 4px;color:#7baaff;font-size:.78rem;font-weight:800;letter-spacing:.14em;text-transform:uppercase}h1{margin:0 0 18px;color:#f4f8ff;font-size:1.7rem;letter-spacing:-.03em}label{display:block;margin:14px 0 5px;color:#eaf2ff;font-weight:750}input{width:100%;border:1px solid #42617f;border-radius:9px;padding:10px 11px;background:#0a1626;color:#e5efff;font:inherit}input:focus{outline:2px solid #5d97f5;outline-offset:1px}.button{display:block;width:100%;margin-top:22px;border:0;border-radius:10px;padding:12px 16px;background:#67e3b8;color:#062018;font:inherit;font-weight:850;cursor:pointer}.error{margin:14px 0 0;color:#ffacba}.alt{margin:18px 0 0;color:#9fb3cf;font-size:.92rem}</style>
+</head>
+<body>
+  <main class="card">
+    <p class="eyebrow">SciMesh</p>
+    <h1>Sign in</h1>
+    <form method="post" action="/ui/login">
+      <label for="email">Email</label>
+      <input id="email" name="email" type="email" autocomplete="username" required autofocus>
+      <label for="password">Password</label>
+      <input id="password" name="password" type="password" autocomplete="current-password" required>
+      <button class="button" type="submit">Sign in</button>
+    </form>
+    {{if .Error}}<p class="error">{{.Error}}</p>{{end}}
+    <p class="alt">No account? <a href="/ui/register">Register</a></p>
+  </main>
+</body>
+</html>
+{{end}}

coordinator/internal/transport/http/templates/new-job.html

@@ -4,27 +4,20 @@
 <head>
   <meta charset="utf-8">
   <meta name="viewport" content="width=device-width, initial-scale=1">
-  <title>Create a check — SciMesh</title>
+  <title>New similarity search · SciMesh</title>
   <style>
-    :root{color:#172033;background:#f6f8fc;font:16px/1.5 system-ui,sans-serif}body{margin:0}.page{max-width:760px;margin:auto;padding:32px 20px 56px}a{color:#174ecf}.back{text-decoration:none}.eyebrow{margin:28px 0 4px;color:#50617d;font-size:.86rem;font-weight:700;text-transform:uppercase;letter-spacing:.08em}h1{margin:0;font-size:2rem}.lead{color:#56657c}.notice{margin:22px 0;padding:16px 18px;border:1px solid #f2cb72;border-radius:10px;background:#fff8e6}.notice strong{display:block}.card{padding:22px;border:1px solid #dfe5f0;border-radius:12px;background:#fff}label{display:block;margin:18px 0 4px;font-weight:700}input{box-sizing:border-box;width:100%;padding:10px;border:1px solid #bac5d8;border-radius:7px;font:inherit}input[type=file]{padding:8px;background:#f8faff}.hint{margin:4px 0;color:#68758b;font-size:.9rem}.button{margin-top:22px;border:0;border-radius:8px;padding:11px 16px;background:#1f5eff;color:#fff;font:inherit;font-weight:700;cursor:pointer}.button:disabled{opacity:.6;cursor:wait}.error{margin-top:16px;color:#a31135}.working{margin-top:16px;color:#174ecf}.checklist{margin:8px 0;padding-left:20px;color:#56657c}.checklist li{margin:5px 0}
+    :root{color:#e5efff;background:#08111f;font:16px/1.5 Inter,ui-sans-serif,system-ui,sans-serif;color-scheme:dark}*{box-sizing:border-box}body{margin:0;background:radial-gradient(circle at 10% -8%,#183f77 0,transparent 32rem),#08111f}.page{max-width:980px;margin:auto;padding:28px 22px 64px}a{color:#94bdff}.back{text-decoration:none}.eyebrow{margin:28px 0 4px;color:#7baaff;font-size:.78rem;font-weight:800;letter-spacing:.14em;text-transform:uppercase}h1{margin:0;color:#f4f8ff;font-size:clamp(2rem,5vw,3.25rem);letter-spacing:-.055em}.lead{max-width:720px;margin:10px 0 0;color:#aabed9;font-size:1.06rem}.layout{display:grid;grid-template-columns:1.45fr .8fr;gap:15px;margin-top:28px}.card,.aside,.notice{border:1px solid #294662;border-radius:15px;background:#0d1a2cdd;box-shadow:0 20px 45px #00000021}.card{padding:22px}.aside,.notice{padding:18px}.aside h2,.notice h2{margin:0;color:#f1f6ff;font-size:1rem}.aside p,.notice p{color:#9fb3cf}.aside ol{margin:13px 0 0;padding-left:20px;color:#aebfda}.aside li{margin:10px 0}label{display:block;margin:18px 0 5px;color:#eaf2ff;font-weight:750}input,select{width:100%;border:1px solid #42617f;border-radius:9px;padding:10px 11px;background:#0a1626;color:#e5efff;font:inherit}input:focus,select:focus{outline:2px solid #5d97f5;outline-offset:1px}input[type=file]{padding:8px}.hint{margin:5px 0;color:#92a9c6;font-size:.87rem}.split{display:grid;grid-template-columns:1fr 1fr;gap:12px}.run-preview{margin-top:18px;border:1px solid #2c8c70;border-radius:11px;padding:13px;background:#0c2b2a;color:#a8f1d0}.run-preview strong{color:#e6fff4}.button{display:inline-flex;margin-top:22px;border:0;border-radius:10px;padding:12px 16px;background:#67e3b8;color:#062018;font:inherit;font-weight:850;cursor:pointer}.button:disabled{opacity:.6;cursor:wait}.working{margin:14px 0 0;color:#9fc5ff}.error{margin:12px 0 0;color:#ffacba}.hidden{display:none}.cap{display:inline-block;margin:4px 5px 0 0;border:1px solid #375978;border-radius:5px;padding:2px 6px;color:#b5d3f5;font:.76rem ui-monospace,SFMono-Regular,monospace}@media(max-width:720px){.layout,.split{grid-template-columns:1fr}.page{padding:22px 14px}}
   </style>
 </head>
 <body>
   <main class="page">
-    <a class="back" href="/ui">← Back to jobs</a><p class="eyebrow">Guided run</p><h1>Search for similar molecules</h1><p class="lead">Creates a diagnostic <code>similarity-search</code> job: a worker finds the top-k molecules most similar to a target SMILES.</p>
-    <section class="notice"><strong>Before starting</strong><ul class="checklist"><li>Keep at least one <code>scimesh-worker</code> running.</li><li>Use a small TSV for a hands-on check.</li><li><b>“Rows per shard” does not limit the file size.</b> It splits the entire upload into tasks: a full ChEMBL TSV at 1,000 rows per shard creates thousands of tasks.</li></ul></section>
-    <form id="run" class="card">
-      <label for="file">ChEMBL TSV</label><input id="file" type="file" name="file" required accept=".tsv,.txt,text/tab-separated-values"><p class="hint">Expected columns: <code>chembl_id</code> and <code>canonical_smiles</code>.</p>
-      <label for="query-smiles">Target molecule (SMILES)</label><input id="query-smiles" name="query_smiles" required maxlength="200" value="CCO" autocomplete="off"><p class="hint"><code>CCO</code> is ethanol. For gefitinib, use its SMILES here or the local CLI with <code>--query-id</code>.</p>
-      <label for="top-k">Matches to return</label><input id="top-k" name="top_k" type="number" min="1" max="100000" value="20" required><p class="hint">This is the top-k within each shard, not a global top-k for the whole dataset yet.</p>
-      <label for="chunk-rows">Rows per shard</label><input id="chunk-rows" name="chunk_rows" type="number" min="1" max="100000" value="1000" required><p class="hint">Fewer rows mean more tasks and more visible progress; more rows mean fewer, longer tasks.</p>
-      <label for="max-rows">Maximum dataset rows to process <small>(optional)</small></label><input id="max-rows" name="max_rows" type="number" min="1" max="10000000" placeholder="For example: 500"><p class="hint">Useful for a quick check of a large TSV. The coordinator creates shards from only the first N data rows; it still stores the original upload.</p>
-      <button class="button" id="submit" type="submit">Upload file and create job</button><p id="working" class="working" hidden aria-live="polite">Uploading the file and creating shard tasks… Keep this page open.</p><p id="error" class="error" role="alert"></p>
-    </form>
+    <a class="back" href="/ui">← Back to control room</a><p class="eyebrow">New computation</p><h1>Similarity search, end to end</h1><p class="lead">Upload a ChEMBL-style TSV. Workers calculate shard candidates; after every shard succeeds, SciMesh merges the exact global top-k into one final CSV.</p>
+    <div class="layout"><form id="run" class="card" novalidate><label for="file">ChEMBL TSV</label><input id="file" type="file" name="file" required accept=".tsv,.txt,text/tab-separated-values"><p class="hint">Required columns: <code>chembl_id</code> and <code>canonical_smiles</code>.</p><label for="query-smiles">Target molecule (SMILES)</label><input id="query-smiles" name="query_smiles" required maxlength="200" value="CCO" autocomplete="off"><p class="hint">Use a valid SMILES. The coordinator shares this exact query with every shard.</p><div class="split"><div><label for="top-k">Global top-k</label><input id="top-k" name="top_k" type="number" min="1" max="100000" value="20" required><p class="hint">How many final molecules to retain.</p></div><div><label for="chunk-rows">Rows per shard</label><input id="chunk-rows" name="chunk_rows" type="number" min="1" max="100000" value="1000" required><p class="hint">Smaller shards make more visible tasks.</p></div></div><div class="split"><div><label for="threshold">Similarity threshold <small>(optional)</small></label><input id="threshold" name="threshold" type="number" min="0" max="1" step="0.01" placeholder="For example: 0.70"><p class="hint">Leave blank to rank every valid candidate.</p></div><div><label for="direction">Keep molecules</label><select id="direction" name="threshold_direction"><option value="greater">more similar (≥ threshold)</option><option value="less">less similar (≤ threshold)</option></select><p class="hint">“Less” helps explore dissimilar molecules.</p></div></div><label for="max-rows">Maximum dataset rows <small>(optional quick run)</small></label><input id="max-rows" name="max_rows" type="number" min="1" max="10000000" placeholder="For example: 500"><p class="hint">Only the first N data rows become shards; the original upload remains stored by the coordinator.</p><div id="preview" class="run-preview"><strong>Ready to plan a run.</strong><br>Select a TSV to see the file that will be sent to the coordinator.</div><button class="button" id="submit" type="submit">Create pipeline run →</button><p id="working" class="working hidden" aria-live="polite">Uploading TSV and creating coordinator-owned shards…</p><p id="error" class="error" role="alert"></p></form><aside class="aside"><h2>What you will observe</h2><ol><li><strong>Input accepted</strong><br>TSV is validated and split into durable shard artifacts.</li><li><strong>Workers claim tasks</strong><br>Each worker downloads one shard, fingerprints it, and uploads a partial CSV.</li><li><strong>Global reduction</strong><br>The coordinator compares exact scores from all partial results.</li><li><strong>Final download</strong><br>The result page exposes a checksum-protected global CSV.</li></ol><h2 style="margin-top:26px">Before you submit</h2><p>Keep at least one worker running in another terminal. The browser cannot start or control worker processes.</p><p><span class="cap">similarity-search</span> is currently the only distributed workload available here.</p></aside></div>
   </main>
   <script>
-    const form=document.querySelector('#run'),button=document.querySelector('#submit'),working=document.querySelector('#working'),error=document.querySelector('#error');
-    form.addEventListener('submit',async event=>{event.preventDefault();error.textContent='';const fields=new FormData(form),file=fields.get('file'),maxRows=String(fields.get('max_rows')||'').trim();if(!(file instanceof File)||file.size===0){error.textContent='Choose a non-empty TSV file.';return}const parameters={query_smiles:fields.get('query_smiles'),top_k:Number(fields.get('top_k')),progress_every:0},upload=new FormData();upload.append('workload','similarity-search');upload.append('parameters',JSON.stringify(parameters));upload.append('chunk_rows',fields.get('chunk_rows'));if(maxRows)upload.append('max_rows',maxRows);upload.append('file',file,file.name);button.disabled=true;working.hidden=false;try{const response=await fetch('/ui/api/jobs/upload',{method:'POST',body:upload}),data=await response.json();if(!response.ok)throw Error(data.error||'Unable to create the job.');location.href='/ui/jobs/'+data.job_id}catch(err){error.textContent=err.message==='invalid input'?'Check the TSV and fields: the coordinator could not accept this request.':err.message;button.disabled=false;working.hidden=true}});
+    const form=document.querySelector('#run'),button=document.querySelector('#submit'),working=document.querySelector('#working'),error=document.querySelector('#error'),preview=document.querySelector('#preview'),fileInput=document.querySelector('#file');
+    const bytes=size=>size<1024?size+' B':size<1024*1024?(size/1024).toFixed(1)+' KiB':(size/(1024*1024)).toFixed(1)+' MiB';fileInput.addEventListener('change',()=>{const file=fileInput.files&&fileInput.files[0];preview.replaceChildren();if(!file){preview.append(document.createTextNode('Select a TSV to see the file that will be sent to the coordinator.'));return}const strong=document.createElement('strong');strong.textContent='Source ready: '+file.name;preview.append(strong,document.createElement('br'),document.createTextNode(bytes(file.size)+' · the coordinator will validate its header before creating tasks.'))});
+    form.addEventListener('submit',async event=>{event.preventDefault();error.textContent='';const fields=new FormData(form),file=fields.get('file'),query=String(fields.get('query_smiles')||'').trim(),topK=Number(fields.get('top_k')),chunkRows=Number(fields.get('chunk_rows')),threshold=String(fields.get('threshold')||'').trim(),maxRows=String(fields.get('max_rows')||'').trim();if(!(file instanceof File)||file.size===0){error.textContent='Choose a non-empty TSV file.';return}if(!query||query.length>200||!Number.isInteger(topK)||topK<1||!Number.isInteger(chunkRows)||chunkRows<1){error.textContent='Enter a target SMILES, a positive global top-k, and a positive rows-per-shard value.';return}if(threshold&&(Number.isNaN(Number(threshold))||Number(threshold)<0||Number(threshold)>1)){error.textContent='Similarity threshold must be between 0 and 1.';return}const parameters={query_smiles:query,top_k:topK,threshold_direction:fields.get('threshold_direction'),progress_every:0};if(threshold)parameters.threshold=Number(threshold);const upload=new FormData();upload.append('workload','similarity-search');upload.append('parameters',JSON.stringify(parameters));upload.append('chunk_rows',String(chunkRows));if(maxRows)upload.append('max_rows',maxRows);upload.append('file',file,file.name);button.disabled=true;working.classList.remove('hidden');try{const response=await fetch('/ui/api/jobs/upload',{method:'POST',body:upload}),data=await response.json();if(!response.ok)throw Error(data.error||'Unable to create the job.');location.href='/ui/jobs/'+encodeURIComponent(data.job_id)}catch(err){error.textContent=err.message==='invalid input'?'The coordinator could not accept this run. Check the TSV columns and form values.':err.message;button.disabled=false;working.classList.add('hidden')}});
   </script>
 </body>
 </html>

coordinator/internal/transport/http/templates/profile.html

@@ -0,0 +1,32 @@
+{{define "profile.html"}}
+<!doctype html>
+<html lang="en">
+<head>
+  <meta charset="utf-8">
+  <meta name="viewport" content="width=device-width, initial-scale=1">
+  <title>Profile · SciMesh</title>
+  <style>
+    :root{color:#e5efff;background:#08111f;font:16px/1.5 Inter,ui-sans-serif,system-ui,sans-serif;color-scheme:dark}*{box-sizing:border-box}body{margin:0;background:radial-gradient(circle at 10% -8%,#183f77 0,transparent 32rem),#08111f}.page{max-width:720px;margin:auto;padding:28px 22px 64px}a{color:#94bdff}.top{display:flex;justify-content:space-between;align-items:center;gap:12px}.eyebrow{margin:0;color:#7baaff;font-size:.78rem;font-weight:800;letter-spacing:.14em;text-transform:uppercase}h1{margin:6px 0 0;color:#f4f8ff;font-size:clamp(1.8rem,4vw,2.6rem);letter-spacing:-.04em}.btn{border:0;border-radius:9px;padding:10px 14px;font:inherit;font-weight:800;cursor:pointer;text-decoration:none}.btn-muted{background:#23344d;color:#dce8ff}.card{margin-top:24px;border:1px solid #294662;border-radius:15px;background:#0d1a2cdd;box-shadow:0 20px 45px #00000021;padding:6px 22px}.err{margin-top:24px;border-radius:10px;padding:12px 14px;background:#552334;color:#ff9bad;font-weight:700}.row{display:flex;justify-content:space-between;gap:16px;padding:15px 0;border-bottom:1px solid #1d3350}.row:last-child{border-bottom:0}.k{color:#9fb3cf}.v{color:#f2f7ff;font-weight:700;text-align:right;word-break:break-all}.mono{font-family:ui-monospace,SFMono-Regular,monospace;font-size:.9rem}.pill{display:inline-block;border-radius:999px;padding:3px 10px;font-size:.82rem;font-weight:800}.pill-yes{background:#123f34;color:#76efb5}.pill-no{background:#23344d;color:#b9cce9}.hint{margin-top:14px;color:#8ba2c2;font-size:.86rem}</style>
+</head>
+<body>
+  <main class="page">
+    <header class="top">
+      <div><p class="eyebrow">Account</p><h1>Your profile</h1></div>
+      <div style="display:flex;gap:10px;align-items:center"><a href="/ui">← Dashboard</a><form method="post" action="/ui/logout" style="margin:0"><button class="btn btn-muted" type="submit">Log out</button></form></div>
+    </header>
+
+    {{if .Error}}<div class="err">{{.Error}}</div>{{end}}
+    {{with .Profile}}
+    <section class="card">
+      <div class="row"><span class="k">User id</span><span class="v mono">{{.ID}}</span></div>
+      <div class="row"><span class="k">Email</span><span class="v">{{.Email}}</span></div>
+      <div class="row"><span class="k">Role</span><span class="v">{{.Role}}</span></div>
+      <div class="row"><span class="k">Verified contributor</span><span class="v">{{if .Verified}}<span class="pill pill-yes">yes</span>{{else}}<span class="pill pill-no">no</span>{{end}}</span></div>
+      <div class="row"><span class="k">Member since</span><span class="v mono">{{.CreatedAt}}</span></div>
+    </section>
+    <p class="hint">Your user id is what the coordinator stores as the owner of every job you submit. Give it to an admin to be promoted or verified.</p>
+    {{end}}
+  </main>
+</body>
+</html>
+{{end}}

coordinator/internal/transport/http/templates/register.html

@@ -0,0 +1,28 @@
+{{define "register.html"}}
+<!doctype html>
+<html lang="en">
+<head>
+  <meta charset="utf-8">
+  <meta name="viewport" content="width=device-width, initial-scale=1">
+  <title>Register · SciMesh</title>
+  <style>
+    :root{color:#e5efff;background:#08111f;font:16px/1.5 Inter,ui-sans-serif,system-ui,sans-serif;color-scheme:dark}*{box-sizing:border-box}body{margin:0;min-height:100vh;display:grid;place-items:center;background:radial-gradient(circle at 10% -8%,#183f77 0,transparent 32rem),#08111f}a{color:#94bdff}.card{width:min(92vw,380px);border:1px solid #294662;border-radius:15px;background:#0d1a2cdd;box-shadow:0 20px 45px #00000021;padding:28px}.eyebrow{margin:0 0 4px;color:#7baaff;font-size:.78rem;font-weight:800;letter-spacing:.14em;text-transform:uppercase}h1{margin:0 0 18px;color:#f4f8ff;font-size:1.7rem;letter-spacing:-.03em}label{display:block;margin:14px 0 5px;color:#eaf2ff;font-weight:750}input{width:100%;border:1px solid #42617f;border-radius:9px;padding:10px 11px;background:#0a1626;color:#e5efff;font:inherit}input:focus{outline:2px solid #5d97f5;outline-offset:1px}.hint{margin:5px 0 0;color:#92a9c6;font-size:.85rem}.button{display:block;width:100%;margin-top:22px;border:0;border-radius:10px;padding:12px 16px;background:#67e3b8;color:#062018;font:inherit;font-weight:850;cursor:pointer}.error{margin:14px 0 0;color:#ffacba}.alt{margin:18px 0 0;color:#9fb3cf;font-size:.92rem}</style>
+</head>
+<body>
+  <main class="card">
+    <p class="eyebrow">SciMesh</p>
+    <h1>Create account</h1>
+    <form method="post" action="/ui/register">
+      <label for="email">Email</label>
+      <input id="email" name="email" type="email" autocomplete="username" required autofocus>
+      <label for="password">Password</label>
+      <input id="password" name="password" type="password" autocomplete="new-password" minlength="8" maxlength="72" required>
+      <p class="hint">At least 8 characters.</p>
+      <button class="button" type="submit">Register</button>
+    </form>
+    {{if .Error}}<p class="error">{{.Error}}</p>{{end}}
+    <p class="alt">Already have an account? <a href="/ui/login">Sign in</a></p>
+  </main>
+</body>
+</html>
+{{end}}

coordinator/internal/transport/http/ui.go

@@ -26,6 +26,7 @@ var uiTemplates = template.Must(template.New("ui").Funcs(template.FuncMap{
 	"taskErrorLabel":    uiTaskErrorLabel,
 	"taskErrorHint":     uiTaskErrorHint,
 	"workerStatusLabel": uiWorkerStatusLabel,
+	"workerStatusClass": uiWorkerStatusClass,
 	"workloadLabel":     uiWorkloadLabel,
 	"progressPercent":   uiProgressPercent,
 	"cancellable":       uiCancellable,
@@ -48,8 +49,10 @@ func uiStatusLabel(status string) string {
 		return "Assigned to a worker"
 	case "running":
 		return "Running"
+	case "reducing":
+		return "Merging results"
 	case "completed":
-		return "Tasks complete"
+		return "Completed"
 	case "failed":
 		return "Needs attention"
 	case "cancelled":
@@ -67,8 +70,10 @@ func uiStatusHint(status string) string {
 		return "A worker has claimed the task and should begin processing shortly."
 	case "running":
 		return "A worker is reading a shard, calculating fingerprints, and uploading its result through the coordinator."
+	case "reducing":
+		return "All shards are complete. The coordinator is merging their candidates into one final CSV."
 	case "completed":
-		return "Every shard task is complete. Files below are still partial results."
+		return "The final result is ready to download."
 	case "failed":
 		return "One or more shard tasks failed. Open the task list below for details."
 	case "cancelled":
@@ -86,7 +91,7 @@ func uiStatusClass(status string) string {
 		return "danger"
 	case "cancelled":
 		return "waiting"
-	case "running", "leased":
+	case "running", "leased", "reducing":
 		return "active"
 	default:
 		return "waiting"
@@ -97,6 +102,8 @@ func uiWorkerStatusLabel(status string) string {
 	switch status {
 	case "online":
 		return "Available"
+	case "busy":
+		return "Busy"
 	case "offline":
 		return "Offline"
 	default:
@@ -104,6 +111,17 @@ func uiWorkerStatusLabel(status string) string {
 	}
 }
 
+func uiWorkerStatusClass(status string) string {
+	switch status {
+	case "online":
+		return "success"
+	case "busy":
+		return "active"
+	default:
+		return "waiting"
+	}
+}
+
 // uiTaskErrorLabel deliberately maps worker implementation errors to an
 // operator-facing diagnosis. Raw subprocess commands and local paths belong in
 // the worker terminal, not in the web UI.
@@ -203,6 +221,20 @@ func (s *Server) handleUIHome(w http.ResponseWriter, r *http.Request) {
 	s.renderUI(w, "dashboard.html", view)
 }
 
+// handleUIOverviewJSON is the bounded polling projection used by the operator
+// dashboard. It intentionally returns only the safe UI read model, never
+// worker tokens, storage keys, or database entities.
+func (s *Server) handleUIOverviewJSON(w http.ResponseWriter, r *http.Request) {
+	ctx, cancel := s.reqCtx(r)
+	defer cancel()
+	view, err := s.uc.Dashboard.Overview(ctx, 20)
+	if err != nil {
+		s.writeError(w, r, err)
+		return
+	}
+	writeJSON(w, http.StatusOK, view)
+}
+
 func (s *Server) handleUINewJob(w http.ResponseWriter, r *http.Request) {
 	s.renderUI(w, "new-job.html", nil)
 }
@@ -253,7 +285,7 @@ func (s *Server) handleUIArtifactDownload(w http.ResponseWriter, r *http.Request
 	}
 	ctx, cancel := s.reqCtx(r)
 	defer cancel()
-	belongs, err := s.uc.Dashboard.ArtifactBelongsToJob(ctx, jobID, artifactID)
+	belongs, err := s.uc.Dashboard.DownloadableArtifactBelongsToJob(ctx, jobID, artifactID)
 	if err != nil {
 		s.writeError(w, r, err)
 		return
@@ -279,3 +311,30 @@ func (s *Server) handleUIArtifactDownload(w http.ResponseWriter, r *http.Request
 	w.Header().Set("X-Checksum-SHA256", art.SHA256)
 	_, _ = io.Copy(w, body)
 }
+
+// handleUIArtifactPreview renders a bounded CSV preview. Its use case owns
+// the job-scoped access rule, including the requirement that a final artifact
+// is the persisted result of a completed job.
+func (s *Server) handleUIArtifactPreview(w http.ResponseWriter, r *http.Request) {
+	jobID, ok := s.uiJobID(w, r)
+	if !ok {
+		return
+	}
+	artifactID, err := uuid.Parse(r.PathValue("artifact_id"))
+	if err != nil {
+		s.writeError(w, r, domain.ErrInvalidInput)
+		return
+	}
+	if s.uc.PreviewArtifact == nil {
+		http.NotFound(w, r)
+		return
+	}
+	ctx, cancel := s.reqCtx(r)
+	defer cancel()
+	view, err := s.uc.PreviewArtifact.Execute(ctx, jobID, artifactID)
+	if err != nil {
+		s.writeError(w, r, err)
+		return
+	}
+	s.renderUI(w, "artifact-preview.html", view)
+}

coordinator/internal/transport/http/ui_admin.go

@@ -0,0 +1,114 @@
+package http
+
+import (
+	"context"
+	"io"
+	"net/http"
+	"net/url"
+	"strings"
+
+	"github.com/google/uuid"
+
+	"github.com/emil28092005/SciMesh/coordinator/internal/authctx"
+)
+
+// adminUserActions are the userservice endpoints the admin panel may invoke, by
+// their path suffix. A whitelist so a crafted form can never proxy an arbitrary
+// path.
+var adminUserActions = map[string]bool{
+	"promote":  true,
+	"demote":   true,
+	"verify":   true,
+	"unverify": true,
+}
+
+// requireAdmin gates a route on the session caller being an admin. It runs
+// inside withUISession, which has already stamped the requester. A non-admin is
+// sent back to the dashboard rather than shown the panel.
+func requireAdmin(next http.Handler) http.Handler {
+	return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		if req, ok := authctx.From(r.Context()); !ok || !req.IsAdmin() {
+			http.Redirect(w, r, "/ui", http.StatusSeeOther)
+			return
+		}
+		next.ServeHTTP(w, r)
+	})
+}
+
+func (s *Server) handleUIAdmin(w http.ResponseWriter, r *http.Request) {
+	role := ""
+	if req, ok := authctx.From(r.Context()); ok {
+		role = req.Role
+	}
+	s.renderUI(w, "admin.html", map[string]any{
+		"Role":  role,
+		"Msg":   r.URL.Query().Get("msg"),
+		"Error": r.URL.Query().Get("error"),
+	})
+}
+
+// handleUIAdminUserAction proxies a user-management action to the userservice,
+// forwarding the admin's session token so the userservice re-checks the role.
+// The user id and action come from the form, so a single static form action can
+// drive every operation.
+func (s *Server) handleUIAdminUserAction(w http.ResponseWriter, r *http.Request) {
+	userID := strings.TrimSpace(r.FormValue("user_id"))
+	action := r.FormValue("action")
+
+	if !adminUserActions[action] {
+		http.Redirect(w, r, "/ui/admin?error=unknown+action", http.StatusSeeOther)
+		return
+	}
+	if _, err := uuid.Parse(userID); err != nil {
+		http.Redirect(w, r, "/ui/admin?error=invalid+user+id", http.StatusSeeOther)
+		return
+	}
+	c, err := r.Cookie(sessionCookie)
+	if err != nil {
+		redirectToLogin(w, r)
+		return
+	}
+
+	status, _, err := s.callUserserviceAuthed(r.Context(), http.MethodPost, "/users/"+userID+"/"+action, c.Value)
+	if err != nil {
+		s.log.Error("admin action proxy", "err", err, "action", action)
+		http.Redirect(w, r, "/ui/admin?error=service+unavailable", http.StatusSeeOther)
+		return
+	}
+	switch status {
+	case http.StatusNoContent:
+		http.Redirect(w, r, "/ui/admin?msg="+url.QueryEscape(action+" applied"), http.StatusSeeOther)
+	case http.StatusNotFound:
+		http.Redirect(w, r, "/ui/admin?error=user+not+found", http.StatusSeeOther)
+	case http.StatusForbidden, http.StatusUnauthorized:
+		http.Redirect(w, r, "/ui/admin?error=not+authorized", http.StatusSeeOther)
+	default:
+		http.Redirect(w, r, "/ui/admin?error=action+failed", http.StatusSeeOther)
+	}
+}
+
+// callUserserviceAuthed makes an authenticated call to the userservice, passing
+// the caller's JWT through as a bearer token. Used for admin actions; login and
+// registration use the unauthenticated callUserservice.
+func (s *Server) callUserserviceAuthed(ctx context.Context, method, path, bearer string) (int, []byte, error) {
+	// path is not attacker-controlled: the caller composes it only from a
+	// uuid-validated id and an action from a fixed whitelist, and the host is
+	// the operator-configured userservice — so the SSRF taint gosec sees here
+	// cannot reach an arbitrary destination.
+	req, err := http.NewRequestWithContext(ctx, method, s.userserviceURL+path, nil) //nolint:gosec // G704: path is validated, host is config
+	if err != nil {
+		return 0, nil, err
+	}
+	req.Header.Set("Authorization", "Bearer "+bearer)
+
+	resp, err := s.httpClient.Do(req) //nolint:gosec // G704: see above
+	if err != nil {
+		return 0, nil, err
+	}
+	defer func() { _ = resp.Body.Close() }()
+	body, err := io.ReadAll(io.LimitReader(resp.Body, 1<<20))
+	if err != nil {
+		return 0, nil, err
+	}
+	return resp.StatusCode, body, nil
+}

coordinator/internal/transport/http/ui_admin_internal_test.go

@@ -0,0 +1,113 @@
+package http
+
+import (
+	"context"
+	"net/http"
+	"net/http/httptest"
+	"net/url"
+	"strings"
+	"testing"
+
+	"github.com/google/uuid"
+
+	"github.com/emil28092005/SciMesh/coordinator/internal/authctx"
+	"github.com/emil28092005/SciMesh/coordinator/internal/usecase"
+)
+
+func adminReq(t *testing.T, role string) *http.Request {
+	t.Helper()
+	req := newReq(http.MethodGet, "/ui/admin", nil)
+	return req.WithContext(authctx.With(context.Background(), authctx.Requester{UserID: uuid.New(), Role: role}))
+}
+
+func TestRequireAdminAllowsAdminOnly(t *testing.T) {
+	reached := false
+	h := requireAdmin(http.HandlerFunc(func(http.ResponseWriter, *http.Request) { reached = true }))
+
+	// Admin passes through.
+	h.ServeHTTP(httptest.NewRecorder(), adminReq(t, "admin"))
+	if !reached {
+		t.Error("admin must reach the handler")
+	}
+
+	// Plain user is redirected to the dashboard.
+	reached = false
+	rec := httptest.NewRecorder()
+	h.ServeHTTP(rec, adminReq(t, "user"))
+	if reached {
+		t.Error("non-admin must not reach the handler")
+	}
+	if rec.Code != http.StatusSeeOther || rec.Header().Get("Location") != "/ui" {
+		t.Errorf("non-admin got %d -> %q, want 303 -> /ui", rec.Code, rec.Header().Get("Location"))
+	}
+}
+
+func TestAdminUserActionForwardsBearer(t *testing.T) {
+	targetID := uuid.NewString()
+	var gotAuth, gotPath string
+	stub := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		gotAuth = r.Header.Get("Authorization")
+		gotPath = r.URL.Path
+		w.WriteHeader(http.StatusNoContent)
+	}))
+	defer stub.Close()
+	s := newLoginServer(stub)
+
+	req := newReq(http.MethodPost, "/ui/admin/user-action",
+		strings.NewReader(url.Values{"user_id": {targetID}, "action": {"promote"}}.Encode()))
+	req.Header.Set("Content-Type", "application/x-www-form-urlencoded")
+	req.AddCookie(&http.Cookie{Name: sessionCookie, Value: "admin.jwt.token"})
+	rec := httptest.NewRecorder()
+	s.handleUIAdminUserAction(rec, req)
+
+	if gotAuth != "Bearer admin.jwt.token" {
+		t.Errorf("forwarded auth = %q, want the admin bearer", gotAuth)
+	}
+	if gotPath != "/users/"+targetID+"/promote" {
+		t.Errorf("forwarded path = %q", gotPath)
+	}
+	if rec.Code != http.StatusSeeOther || !strings.Contains(rec.Header().Get("Location"), "msg=") {
+		t.Errorf("got %d -> %q, want 303 with a success msg", rec.Code, rec.Header().Get("Location"))
+	}
+}
+
+func TestAdminUserActionRejectsUnknownAction(t *testing.T) {
+	s := newLoginServer(httptest.NewServer(http.HandlerFunc(func(http.ResponseWriter, *http.Request) {
+		t.Fatal("userservice must not be called for an invalid action")
+	})))
+	req := newReq(http.MethodPost, "/ui/admin/user-action",
+		strings.NewReader(url.Values{"user_id": {uuid.NewString()}, "action": {"delete"}}.Encode()))
+	req.Header.Set("Content-Type", "application/x-www-form-urlencoded")
+	req.AddCookie(&http.Cookie{Name: sessionCookie, Value: "x"})
+	rec := httptest.NewRecorder()
+	s.handleUIAdminUserAction(rec, req)
+	if !strings.Contains(rec.Header().Get("Location"), "error=") {
+		t.Errorf("unknown action redirect = %q, want an error", rec.Header().Get("Location"))
+	}
+}
+
+func TestAdminUserActionRejectsBadID(t *testing.T) {
+	s := newLoginServer(httptest.NewServer(http.HandlerFunc(func(http.ResponseWriter, *http.Request) {
+		t.Fatal("userservice must not be called for an invalid id")
+	})))
+	req := newReq(http.MethodPost, "/ui/admin/user-action",
+		strings.NewReader(url.Values{"user_id": {"not-a-uuid"}, "action": {"promote"}}.Encode()))
+	req.Header.Set("Content-Type", "application/x-www-form-urlencoded")
+	req.AddCookie(&http.Cookie{Name: sessionCookie, Value: "x"})
+	rec := httptest.NewRecorder()
+	s.handleUIAdminUserAction(rec, req)
+	if !strings.Contains(rec.Header().Get("Location"), "error=") {
+		t.Errorf("bad id redirect = %q, want an error", rec.Header().Get("Location"))
+	}
+}
+
+func TestDashboardAdminLinkOnlyForAdmin(t *testing.T) {
+	admin := render(t, "dashboard.html", usecase.DashboardView{Session: &usecase.SessionView{Role: "admin"}})
+	if !strings.Contains(admin, "/ui/admin") {
+		t.Error("admin must see the Admin link")
+	}
+	user := render(t, "dashboard.html", usecase.DashboardView{Session: &usecase.SessionView{Role: "user"}})
+	if strings.Contains(user, "/ui/admin") {
+		t.Error("a plain user must not see the Admin link")
+	}
+}

coordinator/internal/transport/http/ui_auth.go

@@ -0,0 +1,171 @@
+package http
+
+import (
+	"bytes"
+	"context"
+	"encoding/json"
+	"io"
+	"net/http"
+	"time"
+
+	"github.com/emil28092005/SciMesh/coordinator/internal/authctx"
+	tokenpkg "github.com/emil28092005/SciMesh/coordinator/internal/token"
+)
+
+// sessionCookie holds the userservice JWT for the operator UI. It is httpOnly so
+// page scripts cannot read the token, and scoped to /ui so it never rides along
+// with worker API calls.
+const sessionCookie = "scimesh_session"
+
+// withUISession gates the operator UI on a valid userservice session cookie.
+// A missing or invalid token redirects to the login page rather than returning
+// 401, because the caller here is a browser, not an API client. On success it
+// stamps the requester so downstream handlers can scope views by owner.
+func withUISession(v tokenVerifier) func(http.Handler) http.Handler {
+	return func(next http.Handler) http.Handler {
+		return http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+			c, err := r.Cookie(sessionCookie)
+			if err != nil || c.Value == "" {
+				redirectToLogin(w, r)
+				return
+			}
+			claims, err := v.Verify(c.Value)
+			if err != nil {
+				// Expired or tampered: drop the stale cookie and re-authenticate.
+				clearSessionCookie(w, r)
+				redirectToLogin(w, r)
+				return
+			}
+			ctx := authctx.With(r.Context(), authctx.Requester{
+				UserID:   claims.UserID,
+				Role:     claims.Role,
+				Verified: claims.Verified,
+			})
+			next.ServeHTTP(w, r.WithContext(ctx))
+		})
+	}
+}
+
+// tokenVerifier is satisfied by *token.Verifier; taking an interface keeps the
+// UI auth testable with a stub.
+type tokenVerifier interface {
+	Verify(raw string) (tokenpkg.Claims, error)
+}
+
+func (s *Server) handleUILoginForm(w http.ResponseWriter, r *http.Request) {
+	s.renderUI(w, "login.html", map[string]any{"Error": r.URL.Query().Get("error")})
+}
+
+func (s *Server) handleUIRegisterForm(w http.ResponseWriter, r *http.Request) {
+	s.renderUI(w, "register.html", map[string]any{"Error": r.URL.Query().Get("error")})
+}
+
+// handleUILogin exchanges the submitted credentials for a userservice token and
+// stores it in the session cookie. The coordinator never sees or stores the
+// password beyond forwarding it once.
+func (s *Server) handleUILogin(w http.ResponseWriter, r *http.Request) {
+	email, password := r.FormValue("email"), r.FormValue("password")
+
+	status, body, err := s.callUserservice(r.Context(), "/login", email, password)
+	if err != nil {
+		s.log.Error("userservice login call", "err", err)
+		http.Redirect(w, r, "/ui/login?error=service+unavailable", http.StatusSeeOther)
+		return
+	}
+	if status != http.StatusOK {
+		http.Redirect(w, r, "/ui/login?error=invalid+email+or+password", http.StatusSeeOther)
+		return
+	}
+
+	var resp struct {
+		Token string `json:"token"`
+	}
+	if err := json.Unmarshal(body, &resp); err != nil || resp.Token == "" {
+		http.Redirect(w, r, "/ui/login?error=service+unavailable", http.StatusSeeOther)
+		return
+	}
+	setSessionCookie(w, r, resp.Token)
+	http.Redirect(w, r, "/ui", http.StatusSeeOther)
+}
+
+// handleUIRegister creates an account through the userservice, then sends the
+// user to the login page. The new account is a plain user until an admin
+// promotes or verifies it.
+func (s *Server) handleUIRegister(w http.ResponseWriter, r *http.Request) {
+	email, password := r.FormValue("email"), r.FormValue("password")
+
+	status, _, err := s.callUserservice(r.Context(), "/register", email, password)
+	if err != nil {
+		s.log.Error("userservice register call", "err", err)
+		http.Redirect(w, r, "/ui/register?error=service+unavailable", http.StatusSeeOther)
+		return
+	}
+	switch status {
+	case http.StatusCreated:
+		http.Redirect(w, r, "/ui/login?error=registered,+please+log+in", http.StatusSeeOther)
+	case http.StatusConflict:
+		http.Redirect(w, r, "/ui/register?error=email+already+registered", http.StatusSeeOther)
+	default:
+		http.Redirect(w, r, "/ui/register?error=invalid+email+or+password", http.StatusSeeOther)
+	}
+}
+
+func (s *Server) handleUILogout(w http.ResponseWriter, r *http.Request) {
+	clearSessionCookie(w, r)
+	http.Redirect(w, r, "/ui/login", http.StatusSeeOther)
+}
+
+// callUserservice POSTs credentials to the userservice and returns its status
+// and body. It is the only runtime dependency on the userservice — login and
+// registration; token verification stays local.
+func (s *Server) callUserservice(ctx context.Context, path, email, password string) (int, []byte, error) {
+	payload, _ := json.Marshal(map[string]string{"email": email, "password": password})
+	req, err := http.NewRequestWithContext(ctx, http.MethodPost, s.userserviceURL+path, bytes.NewReader(payload))
+	if err != nil {
+		return 0, nil, err
+	}
+	req.Header.Set("Content-Type", "application/json")
+
+	resp, err := s.httpClient.Do(req)
+	if err != nil {
+		return 0, nil, err
+	}
+	defer func() { _ = resp.Body.Close() }()
+
+	// Cap the response; login/register bodies are tiny.
+	body, err := io.ReadAll(io.LimitReader(resp.Body, 1<<20))
+	if err != nil {
+		return 0, nil, err
+	}
+	return resp.StatusCode, body, nil
+}
+
+func setSessionCookie(w http.ResponseWriter, r *http.Request, token string) {
+	// Secure is set under TLS; a local demo runs plain HTTP, where forcing
+	// Secure would stop the browser from ever sending the cookie back.
+	http.SetCookie(w, &http.Cookie{ //nolint:gosec // G124: Secure follows r.TLS by design
+		Name:     sessionCookie,
+		Value:    token,
+		Path:     "/ui",
+		HttpOnly: true,
+		Secure:   r.TLS != nil,
+		SameSite: http.SameSiteLaxMode,
+		Expires:  time.Now().Add(24 * time.Hour),
+	})
+}
+
+func clearSessionCookie(w http.ResponseWriter, r *http.Request) {
+	http.SetCookie(w, &http.Cookie{ //nolint:gosec // G124: Secure follows r.TLS by design
+		Name:     sessionCookie,
+		Value:    "",
+		Path:     "/ui",
+		HttpOnly: true,
+		Secure:   r.TLS != nil,
+		SameSite: http.SameSiteLaxMode,
+		MaxAge:   -1,
+	})
+}
+
+func redirectToLogin(w http.ResponseWriter, r *http.Request) {
+	http.Redirect(w, r, "/ui/login", http.StatusSeeOther)
+}

coordinator/internal/transport/http/ui_auth_internal_test.go

@@ -0,0 +1,175 @@
+package http
+
+import (
+	"context"
+	"errors"
+	"io"
+	"log/slog"
+	"net/http"
+	"net/http/httptest"
+	"net/url"
+	"strings"
+	"testing"
+
+	"github.com/google/uuid"
+
+	"github.com/emil28092005/SciMesh/coordinator/internal/authctx"
+	tokenpkg "github.com/emil28092005/SciMesh/coordinator/internal/token"
+)
+
+// newReq builds a request carrying a context, which http.NewRequestWithContext
+// provides on go1.22 (httptest.NewRequestWithContext needs go1.23).
+func newReq(method, target string, body io.Reader) *http.Request {
+	req, err := http.NewRequestWithContext(context.Background(), method, target, body)
+	if err != nil {
+		panic(err)
+	}
+	return req
+}
+
+type stubVerifier struct {
+	claims tokenpkg.Claims
+	err    error
+}
+
+func (s stubVerifier) Verify(string) (tokenpkg.Claims, error) { return s.claims, s.err }
+
+func TestWithUISessionRedirectsWithoutCookie(t *testing.T) {
+	h := withUISession(stubVerifier{})(http.HandlerFunc(func(http.ResponseWriter, *http.Request) {
+		t.Fatal("handler must not run without a session")
+	}))
+	rec := httptest.NewRecorder()
+	h.ServeHTTP(rec, newReq(http.MethodGet, "/ui", nil))
+
+	if rec.Code != http.StatusSeeOther {
+		t.Fatalf("got %d, want 303", rec.Code)
+	}
+	if loc := rec.Header().Get("Location"); loc != "/ui/login" {
+		t.Errorf("redirect = %q, want /ui/login", loc)
+	}
+}
+
+func TestWithUISessionAcceptsValidCookieAndStampsRequester(t *testing.T) {
+	id := uuid.New()
+	verifier := stubVerifier{claims: tokenpkg.Claims{UserID: id, Role: "admin", Verified: true}}
+
+	var gotReq authctx.Requester
+	var ok bool
+	h := withUISession(verifier)(http.HandlerFunc(func(_ http.ResponseWriter, r *http.Request) {
+		gotReq, ok = authctx.From(r.Context())
+	}))
+
+	req := newReq(http.MethodGet, "/ui", nil)
+	req.AddCookie(&http.Cookie{Name: sessionCookie, Value: "valid.jwt"})
+	rec := httptest.NewRecorder()
+	h.ServeHTTP(rec, req)
+
+	if !ok || gotReq.UserID != id || gotReq.Role != "admin" || !gotReq.Verified {
+		t.Errorf("requester = %+v (ok=%v), want id=%v admin verified", gotReq, ok, id)
+	}
+}
+
+func TestWithUISessionClearsInvalidCookie(t *testing.T) {
+	h := withUISession(stubVerifier{err: errors.New("expired")})(http.HandlerFunc(func(http.ResponseWriter, *http.Request) {
+		t.Fatal("handler must not run with an invalid token")
+	}))
+	req := newReq(http.MethodGet, "/ui", nil)
+	req.AddCookie(&http.Cookie{Name: sessionCookie, Value: "stale.jwt"})
+	rec := httptest.NewRecorder()
+	h.ServeHTTP(rec, req)
+
+	if rec.Code != http.StatusSeeOther {
+		t.Fatalf("got %d, want 303", rec.Code)
+	}
+	if c := rec.Result().Cookies(); len(c) == 0 || c[0].MaxAge >= 0 {
+		t.Error("stale cookie must be cleared (MaxAge < 0)")
+	}
+}
+
+// newLoginServer builds a Server whose userservice calls hit stub.
+func newLoginServer(stub *httptest.Server) *Server {
+	return &Server{
+		log:            slog.New(slog.NewTextHandler(io.Discard, nil)),
+		userserviceURL: strings.TrimRight(stub.URL, "/"),
+		httpClient:     stub.Client(),
+	}
+}
+
+func postForm(path string, form url.Values) *http.Request {
+	req := newReq(http.MethodPost, path, strings.NewReader(form.Encode()))
+	req.Header.Set("Content-Type", "application/x-www-form-urlencoded")
+	return req
+}
+
+func TestHandleUILoginSetsCookieOnSuccess(t *testing.T) {
+	stub := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		if r.URL.Path != "/login" {
+			t.Errorf("unexpected path %q", r.URL.Path)
+		}
+		w.WriteHeader(http.StatusOK)
+		_, _ = w.Write([]byte(`{"token":"issued.jwt.here"}`))
+	}))
+	defer stub.Close()
+	s := newLoginServer(stub)
+
+	rec := httptest.NewRecorder()
+	s.handleUILogin(rec, postForm("/ui/login", url.Values{"email": {"a@b.com"}, "password": {"password123"}}))
+
+	if rec.Code != http.StatusSeeOther || rec.Header().Get("Location") != "/ui" {
+		t.Fatalf("got %d -> %q, want 303 -> /ui", rec.Code, rec.Header().Get("Location"))
+	}
+	cookies := rec.Result().Cookies()
+	if len(cookies) == 0 || cookies[0].Name != sessionCookie || cookies[0].Value != "issued.jwt.here" {
+		t.Errorf("session cookie not set: %+v", cookies)
+	}
+	if !cookies[0].HttpOnly {
+		t.Error("session cookie must be httpOnly")
+	}
+}
+
+func TestHandleUILoginRejectsBadCredentials(t *testing.T) {
+	stub := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
+		w.WriteHeader(http.StatusUnauthorized)
+	}))
+	defer stub.Close()
+	s := newLoginServer(stub)
+
+	rec := httptest.NewRecorder()
+	s.handleUILogin(rec, postForm("/ui/login", url.Values{"email": {"a@b.com"}, "password": {"wrong"}}))
+
+	if rec.Code != http.StatusSeeOther || !strings.HasPrefix(rec.Header().Get("Location"), "/ui/login?error=") {
+		t.Fatalf("got %d -> %q, want 303 -> /ui/login?error=", rec.Code, rec.Header().Get("Location"))
+	}
+	if len(rec.Result().Cookies()) != 0 {
+		t.Error("no cookie must be set on failed login")
+	}
+}
+
+func TestHandleUIRegisterConflict(t *testing.T) {
+	stub := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, _ *http.Request) {
+		w.WriteHeader(http.StatusConflict)
+	}))
+	defer stub.Close()
+	s := newLoginServer(stub)
+
+	rec := httptest.NewRecorder()
+	s.handleUIRegister(rec, postForm("/ui/register", url.Values{"email": {"dup@b.com"}, "password": {"password123"}}))
+
+	if got := rec.Header().Get("Location"); !strings.Contains(got, "already+registered") {
+		t.Errorf("register conflict redirect = %q", got)
+	}
+}
+
+func TestHandleUILogoutClearsCookie(t *testing.T) {
+	s := &Server{log: slog.New(slog.NewTextHandler(io.Discard, nil))}
+	rec := httptest.NewRecorder()
+	s.handleUILogout(rec, newReq(http.MethodPost, "/ui/logout", nil))
+
+	if rec.Header().Get("Location") != "/ui/login" {
+		t.Errorf("logout redirect = %q", rec.Header().Get("Location"))
+	}
+	c := rec.Result().Cookies()
+	if len(c) == 0 || c[0].MaxAge >= 0 {
+		t.Error("logout must clear the session cookie")
+	}
+}

coordinator/internal/transport/http/ui_logout_internal_test.go

@@ -0,0 +1,42 @@
+package http
+
+import (
+	"bytes"
+	"strings"
+	"testing"
+
+	"github.com/emil28092005/SciMesh/coordinator/internal/usecase"
+)
+
+func render(t *testing.T, name string, data any) string {
+	t.Helper()
+	var buf bytes.Buffer
+	if err := uiTemplates.ExecuteTemplate(&buf, name, data); err != nil {
+		t.Fatalf("render %s: %v", name, err)
+	}
+	return buf.String()
+}
+
+func TestDashboardLogoutOnlyInSession(t *testing.T) {
+	withSession := render(t, "dashboard.html", usecase.DashboardView{Session: &usecase.SessionView{Role: "admin"}})
+	if !strings.Contains(withSession, "/ui/logout") || !strings.Contains(withSession, "Log out") {
+		t.Error("dashboard must show a logout control in session mode")
+	}
+
+	noSession := render(t, "dashboard.html", usecase.DashboardView{})
+	if strings.Contains(noSession, "/ui/logout") {
+		t.Error("dashboard must not show logout under basic auth (no session)")
+	}
+}
+
+func TestJobLogoutOnlyInSession(t *testing.T) {
+	withSession := render(t, "job.html", usecase.JobDetailView{Session: &usecase.SessionView{Role: "user"}})
+	if !strings.Contains(withSession, "/ui/logout") {
+		t.Error("job page must show a logout control in session mode")
+	}
+
+	noSession := render(t, "job.html", usecase.JobDetailView{})
+	if strings.Contains(noSession, "/ui/logout") {
+		t.Error("job page must not show logout under basic auth (no session)")
+	}
+}

coordinator/internal/transport/http/ui_profile.go

@@ -0,0 +1,49 @@
+package http
+
+import (
+	"encoding/json"
+	"net/http"
+)
+
+// profileView is the account data shown on the profile page, mirroring the
+// userservice /me response.
+type profileView struct {
+	ID        string `json:"id"`
+	Email     string `json:"email"`
+	Role      string `json:"role"`
+	Verified  bool   `json:"verified"`
+	CreatedAt string `json:"created_at"`
+}
+
+// handleUIProfile shows the signed-in user's own account. It proxies the
+// session token to the userservice /me endpoint, which is the authority on the
+// account (email and created_at are not in the JWT).
+func (s *Server) handleUIProfile(w http.ResponseWriter, r *http.Request) {
+	c, err := r.Cookie(sessionCookie)
+	if err != nil {
+		redirectToLogin(w, r)
+		return
+	}
+	status, body, err := s.callUserserviceAuthed(r.Context(), http.MethodGet, "/me", c.Value)
+	if err != nil {
+		s.log.Error("profile /me proxy", "err", err)
+		s.renderUI(w, "profile.html", map[string]any{"Error": "userservice unavailable"})
+		return
+	}
+	if status == http.StatusUnauthorized {
+		clearSessionCookie(w, r)
+		redirectToLogin(w, r)
+		return
+	}
+	if status != http.StatusOK {
+		s.renderUI(w, "profile.html", map[string]any{"Error": "could not load your account"})
+		return
+	}
+
+	var p profileView
+	if err := json.Unmarshal(body, &p); err != nil {
+		s.renderUI(w, "profile.html", map[string]any{"Error": "could not read your account"})
+		return
+	}
+	s.renderUI(w, "profile.html", map[string]any{"Profile": p})
+}

coordinator/internal/transport/http/ui_profile_internal_test.go

@@ -0,0 +1,43 @@
+package http
+
+import (
+	"net/http"
+	"net/http/httptest"
+	"strings"
+	"testing"
+)
+
+func TestProfileProxiesMe(t *testing.T) {
+	var gotAuth, gotPath string
+	stub := httptest.NewServer(http.HandlerFunc(func(w http.ResponseWriter, r *http.Request) {
+		gotAuth, gotPath = r.Header.Get("Authorization"), r.URL.Path
+		w.WriteHeader(http.StatusOK)
+		_, _ = w.Write([]byte(`{"id":"11111111-1111-1111-1111-111111111111","email":"me@example.com","role":"user","verified":false,"created_at":"2026-07-26T00:00:00Z"}`))
+	}))
+	defer stub.Close()
+	s := newLoginServer(stub)
+
+	req := newReq(http.MethodGet, "/ui/profile", nil)
+	req.AddCookie(&http.Cookie{Name: sessionCookie, Value: "my.jwt"})
+	rec := httptest.NewRecorder()
+	s.handleUIProfile(rec, req)
+
+	if gotAuth != "Bearer my.jwt" || gotPath != "/me" {
+		t.Fatalf("proxy: auth=%q path=%q", gotAuth, gotPath)
+	}
+	body := rec.Body.String()
+	if !strings.Contains(body, "me@example.com") || !strings.Contains(body, "11111111-1111-1111-1111-111111111111") {
+		t.Error("profile page must show the email and id")
+	}
+}
+
+func TestProfileRedirectsWithoutCookie(t *testing.T) {
+	s := newLoginServer(httptest.NewServer(http.HandlerFunc(func(http.ResponseWriter, *http.Request) {
+		t.Fatal("must not call userservice without a session")
+	})))
+	rec := httptest.NewRecorder()
+	s.handleUIProfile(rec, newReq(http.MethodGet, "/ui/profile", nil))
+	if rec.Code != http.StatusSeeOther {
+		t.Errorf("no cookie: got %d, want 303 redirect", rec.Code)
+	}
+}

coordinator/internal/transport/http/ui_test.go

@@ -10,7 +10,8 @@ func TestUIStatusPresentation(t *testing.T) {
 	}{
 		{"pending", "Waiting for a worker", "waiting"},
 		{"running", "Running", "active"},
-		{"completed", "Tasks complete", "success"},
+		{"reducing", "Merging results", "active"},
+		{"completed", "Completed", "success"},
 		{"failed", "Needs attention", "danger"},
 	}
 	for _, test := range tests {
@@ -42,3 +43,12 @@ func TestUITaskErrorPresentationDoesNotExposeCommand(t *testing.T) {
 		t.Error("error hint must explain the failure")
 	}
 }
+
+func TestUIWorkerStatusPresentation(t *testing.T) {
+	if got := uiWorkerStatusLabel("busy"); got != "Busy" {
+		t.Errorf("busy worker label = %q", got)
+	}
+	if got := uiWorkerStatusClass("busy"); got != "active" {
+		t.Errorf("busy worker class = %q", got)
+	}
+}

coordinator/internal/usecase/artifact.go

@@ -12,17 +12,22 @@ import (
 // UploadArtifact stores a worker's partial-result bytes and records the metadata.
 type UploadArtifact struct {
 	tasks     TaskRepository
+	workers   WorkerRepository
 	artifacts ArtifactRepository
 	blobs     BlobStore
+	tx        TxManager
 	clk       Clock
 }
 
-func NewUploadArtifact(tasks TaskRepository, artifacts ArtifactRepository,
-	blobs BlobStore, clk Clock) *UploadArtifact {
-	return &UploadArtifact{tasks: tasks, artifacts: artifacts, blobs: blobs, clk: clk}
+func NewUploadArtifact(tasks TaskRepository, workers WorkerRepository, artifacts ArtifactRepository,
+	blobs BlobStore, tx TxManager, clk Clock) *UploadArtifact {
+	return &UploadArtifact{tasks: tasks, workers: workers, artifacts: artifacts, blobs: blobs, tx: tx, clk: clk}
 }
 
 func (uc *UploadArtifact) Execute(ctx context.Context, in UploadArtifactInput) (*domain.Artifact, error) {
+	if err := authorizeWorkerOwner(ctx, uc.workers, in.WorkerID); err != nil {
+		return nil, err
+	}
 	task, err := uc.tasks.Get(ctx, in.TaskID)
 	if err != nil {
 		return nil, err
@@ -32,6 +37,15 @@ func (uc *UploadArtifact) Execute(ctx context.Context, in UploadArtifactInput) (
 	if !task.IsLeaseHeldBy(in.WorkerID, in.Attempt, uc.clk.Now()) {
 		return nil, domain.ErrLeaseConflict
 	}
+	// A client can retry a PUT after losing the response. Return the one durable
+	// result for this lease attempt instead of storing duplicate artifacts.
+	existing, err := uc.artifacts.FindPartialResult(ctx, in.TaskID, in.Attempt)
+	if err != nil {
+		return nil, err
+	}
+	if existing != nil {
+		return existing, nil
+	}
 
 	taskID := task.ID
 	art, err := domain.NewArtifact(task.JobID, &taskID, domain.ArtifactPartialResult,
@@ -50,25 +64,41 @@ func (uc *UploadArtifact) Execute(ctx context.Context, in UploadArtifactInput) (
 	}
 	art.SetContent(sum, size)
 
-	// The stream may take longer than the lease. Re-check after it finishes so
-	// an expired worker cannot leave a durable result record behind. Completion
-	// performs the same ownership check under its transaction.
-	current, err := uc.tasks.Get(ctx, in.TaskID)
+	// The stream may take longer than the lease. Lock the task while re-checking
+	// ownership and inserting metadata: completion or another upload cannot race
+	// this final decision. The database unique index is a second line of defence.
+	var durable *domain.Artifact
+	err = uc.tx.WithinTx(ctx, func(ctx context.Context) error {
+		current, err := uc.tasks.GetForUpdate(ctx, in.TaskID)
+		if err != nil {
+			return err
+		}
+		if !current.IsLeaseHeldBy(in.WorkerID, in.Attempt, uc.clk.Now()) {
+			return domain.ErrLeaseConflict
+		}
+		existing, err := uc.artifacts.FindPartialResult(ctx, in.TaskID, in.Attempt)
+		if err != nil {
+			return err
+		}
+		if existing != nil {
+			durable = existing
+			return nil
+		}
+		if err := uc.artifacts.Insert(ctx, art); err != nil {
+			return err
+		}
+		durable = art
+		return nil
+	})
 	if err != nil {
 		_ = uc.blobs.Delete(ctx, art.StorageKey)
 		return nil, err
 	}
-	if !current.IsLeaseHeldBy(in.WorkerID, in.Attempt, uc.clk.Now()) {
+	if durable != art {
+		// Another request won the race while this stream was being written.
 		_ = uc.blobs.Delete(ctx, art.StorageKey)
-		return nil, domain.ErrLeaseConflict
 	}
-
-	// Persist the record. If that fails the blob would be an orphan, so remove it.
-	if err := uc.artifacts.Insert(ctx, art); err != nil {
-		_ = uc.blobs.Delete(ctx, art.StorageKey)
-		return nil, err
-	}
-	return art, nil
+	return durable, nil
 }
 
 // DownloadArtifact returns an artifact's metadata together with a reader over

coordinator/internal/usecase/dto.go

@@ -4,6 +4,8 @@ import (
 	"io"
 
 	"github.com/google/uuid"
+
+	"github.com/emil28092005/SciMesh/coordinator/internal/domain"
 )
 
 // Use-case boundary types. Adapters map their wire formats onto these, so the
@@ -28,6 +30,11 @@ type ChunkInput struct {
 type RegisterWorkerInput struct {
 	Name         string
 	Capabilities []string
+	// OwnerID is the userservice user registering this worker; nil for a
+	// shared-token registration. TrustLevel is resolved by the transport layer
+	// from how the caller authenticated.
+	OwnerID    *uuid.UUID
+	TrustLevel domain.WorkerTrust
 }
 
 type ClaimTaskInput struct {

coordinator/internal/usecase/job.go

@@ -33,6 +33,17 @@ func NewCreateJob(jobs JobRepository, tasks TaskRepository, tx TxManager, clock
 // The all-or-none guarantee comes from TxManager: a half-created job would
 // leave chunks no worker could ever complete.
 func (uc *CreateJob) Execute(ctx context.Context, in CreateJobInput) (*domain.Job, error) {
+	if in.Workload == "similarity-graph" || in.Workload == "similarity_graph" {
+		// CTX-10 must plan triangular block pairs; ordinary independent input
+		// chunks would silently omit every cross-chunk molecular pair.
+		return nil, domain.ErrInvalidInput
+	}
+	if (in.Workload == "similarity-search" || in.Workload == "similarity_search") &&
+		len(in.Chunks) > 1 && in.Parameters["query_id"] != nil {
+		// Resolving once against the source dataset belongs to CTX-07. Letting
+		// each shard resolve it would make most tasks fail or use inconsistent data.
+		return nil, domain.ErrInvalidInput
+	}
 	chunks := make([]domain.ChunkSpec, 0, len(in.Chunks))
 	for _, c := range in.Chunks {
 		chunks = append(chunks, domain.ChunkSpec(c))
@@ -42,6 +53,7 @@ func (uc *CreateJob) Execute(ctx context.Context, in CreateJobInput) (*domain.Jo
 	if err != nil {
 		return nil, err
 	}
+	job.OwnerID = ownerFromContext(ctx)
 
 	err = uc.tx.WithinTx(ctx, func(ctx context.Context) error {
 		if err := uc.jobs.Insert(ctx, job); err != nil {
@@ -86,10 +98,24 @@ func (uc *CancelJob) Execute(ctx context.Context, jobID uuid.UUID) (int64, error
 		if err != nil {
 			return err
 		}
+		if err := authorizeJobAccess(ctx, job); err != nil {
+			return err
+		}
 		if job.Status == domain.JobCancelled {
 			return nil
 		}
-		if job.Status == domain.JobCompleted || job.Status == domain.JobFailed {
+		if job.Status == domain.JobReducing || job.Status == domain.JobCompleted || job.Status == domain.JobFailed {
+			return domain.ErrJobNotCancellable
+		}
+		// The lease reaper can be the transition that exhausted the final task.
+		// Check the authoritative task histogram as well as the cached job status,
+		// so a stale status can never turn a failed/completed job into cancelled.
+		counts, err := uc.tasks.CountByStatus(ctx, jobID)
+		if err != nil {
+			return err
+		}
+		derived := progressFrom(*job, counts).DeriveStatus()
+		if derived == domain.JobReducing || derived == domain.JobCompleted || derived == domain.JobFailed {
 			return domain.ErrJobNotCancellable
 		}
 		cancelled, err = uc.tasks.CancelByJob(ctx, jobID, now)
@@ -110,6 +136,9 @@ func (uc *GetJobStatus) Execute(ctx context.Context, jobID uuid.UUID) (domain.Jo
 	if err != nil {
 		return domain.JobProgress{}, err
 	}
+	if err := authorizeJobAccess(ctx, job); err != nil {
+		return domain.JobProgress{}, err
+	}
 	counts, err := uc.tasks.CountByStatus(ctx, jobID)
 	if err != nil {
 		return domain.JobProgress{}, err
@@ -204,11 +233,36 @@ func syncJobStatus(ctx context.Context, jobs JobRepository, tasks TaskRepository
 	if err != nil {
 		return err
 	}
-	status := progressFrom(domain.Job{}, counts).DeriveStatus()
+	job, err := jobs.Get(ctx, jobID)
+	if err != nil {
+		return err
+	}
+	status := progressFrom(*job, counts).DeriveStatus()
+	// All worker shards being complete means scientific reduction is ready, not
+	// that the job's final artifact already exists. CTX-09 owns the transition
+	// from reducing to completed after it persists that artifact.
+	if status == domain.JobCompleted && job.Workload == "similarity-search" {
+		status = domain.JobReducing
+	}
 
 	var completedAt *time.Time
-	if status == domain.JobCompleted || status == domain.JobFailed {
+	if status == domain.JobFailed {
 		completedAt = &now
 	}
 	return jobs.UpdateStatus(ctx, jobID, status, completedAt)
 }
+
+func syncExpiredJobStatuses(ctx context.Context, jobs JobRepository, tasks TaskRepository,
+	jobIDs []uuid.UUID, now time.Time) error {
+	seen := make(map[uuid.UUID]struct{}, len(jobIDs))
+	for _, jobID := range jobIDs {
+		if _, duplicate := seen[jobID]; duplicate {
+			continue
+		}
+		seen[jobID] = struct{}{}
+		if err := syncJobStatus(ctx, jobs, tasks, jobID, now); err != nil {
+			return err
+		}
+	}
+	return nil
+}

coordinator/internal/usecase/ownership.go

@@ -0,0 +1,51 @@
+package usecase
+
+import (
+	"context"
+
+	"github.com/google/uuid"
+
+	"github.com/emil28092005/SciMesh/coordinator/internal/authctx"
+	"github.com/emil28092005/SciMesh/coordinator/internal/domain"
+)
+
+// ownerFromContext returns the authenticated user id to stamp on a new job, or
+// nil when the request was not authenticated as a user — worker or legacy
+// traffic, or user-JWT auth disabled. A nil owner is stored as NULL.
+func ownerFromContext(ctx context.Context) *uuid.UUID {
+	if r, ok := authctx.From(ctx); ok {
+		id := r.UserID
+		return &id
+	}
+	return nil
+}
+
+// uiOwnerFilter returns the owner a UI listing must be restricted to: nil for an
+// operator/admin or an unauthenticated (basic-auth) session, which see all jobs,
+// or the caller's id for a plain user, who sees only their own.
+func uiOwnerFilter(ctx context.Context) *uuid.UUID {
+	r, ok := authctx.From(ctx)
+	if !ok || r.IsAdmin() {
+		return nil
+	}
+	id := r.UserID
+	return &id
+}
+
+// authorizeJobAccess enforces that a non-admin user may only act on their own
+// job. It returns ErrJobNotFound — not a 403 — on a mismatch, so the response
+// never reveals that another user's job exists.
+//
+// Requests with no authenticated user (worker/legacy traffic, or JWT auth
+// disabled) are not restricted here: the shared service token already gated
+// them, and worker endpoints legitimately operate across all jobs.
+func authorizeJobAccess(ctx context.Context, job *domain.Job) error {
+	r, ok := authctx.From(ctx)
+	if !ok || r.IsAdmin() {
+		return nil
+	}
+	if job.OwnerID == nil || *job.OwnerID != r.UserID {
+		return domain.ErrJobNotFound
+	}
+	return nil
+}

coordinator/internal/usecase/ownership_test.go

@@ -0,0 +1,66 @@
+package usecase
+
+import (
+	"context"
+	"errors"
+	"testing"
+
+	"github.com/google/uuid"
+
+	"github.com/emil28092005/SciMesh/coordinator/internal/authctx"
+	"github.com/emil28092005/SciMesh/coordinator/internal/domain"
+)
+
+func TestOwnerFromContext(t *testing.T) {
+	if ownerFromContext(context.Background()) != nil {
+		t.Error("no requester must yield a nil owner")
+	}
+	id := uuid.New()
+	ctx := authctx.With(context.Background(), authctx.Requester{UserID: id, Role: "user"})
+	got := ownerFromContext(ctx)
+	if got == nil || *got != id {
+		t.Errorf("owner = %v, want %v", got, id)
+	}
+}
+
+func TestAuthorizeJobAccess(t *testing.T) {
+	owner := uuid.New()
+	other := uuid.New()
+	job := &domain.Job{ID: uuid.New(), OwnerID: &owner}
+
+	ctxOf := func(id uuid.UUID, role string) context.Context {
+		return authctx.With(context.Background(), authctx.Requester{UserID: id, Role: role})
+	}
+
+	cases := []struct {
+		name    string
+		ctx     context.Context
+		wantErr bool
+	}{
+		{"no requester (worker/legacy) allowed", context.Background(), false},
+		{"owner allowed", ctxOf(owner, "user"), false},
+		{"admin allowed", ctxOf(other, "admin"), false},
+		{"non-owner denied", ctxOf(other, "user"), true},
+	}
+	for _, tc := range cases {
+		t.Run(tc.name, func(t *testing.T) {
+			err := authorizeJobAccess(tc.ctx, job)
+			if tc.wantErr {
+				if !errors.Is(err, domain.ErrJobNotFound) {
+					t.Errorf("got %v, want ErrJobNotFound", err)
+				}
+			} else if err != nil {
+				t.Errorf("unexpected error: %v", err)
+			}
+		})
+	}
+}
+
+func TestAuthorizeJobAccessNilOwner(t *testing.T) {
+	// A legacy job with no owner must not be readable by an arbitrary user.
+	job := &domain.Job{ID: uuid.New(), OwnerID: nil}
+	ctx := authctx.With(context.Background(), authctx.Requester{UserID: uuid.New(), Role: "user"})
+	if err := authorizeJobAccess(ctx, job); !errors.Is(err, domain.ErrJobNotFound) {
+		t.Errorf("got %v, want ErrJobNotFound", err)
+	}
+}

coordinator/internal/usecase/ports.go

@@ -23,6 +23,15 @@ type ClaimFilter struct {
 	Owner      string   // worker ID taking the lease
 	Now        time.Time
 	LeaseUntil time.Time
+	// VoterOwner, when set, excludes tasks this owner has already voted on, so
+	// an untrusted worker never verifies its own chunk twice.
+	VoterOwner *uuid.UUID
+}
+
+// TaskResultRepository records and tallies quorum votes for untrusted results.
+type TaskResultRepository interface {
+	RecordVote(ctx context.Context, taskID, ownerID uuid.UUID, sha256 string, artifactID uuid.UUID) error
+	CountAgreeing(ctx context.Context, taskID uuid.UUID, sha256 string) (int, error)
 }
 
 // TaskRepository persists tasks.
@@ -60,9 +69,9 @@ type TaskRepository interface {
 	// lease. It returns how many tasks changed.
 	CancelByJob(ctx context.Context, jobID uuid.UUID, now time.Time) (int64, error)
 
-	// ExpireLeases applies the lease-expiry rule to every elapsed task and
-	// reports how many were affected.
-	ExpireLeases(ctx context.Context, now time.Time) (int64, error)
+	// ExpireLeases applies the lease-expiry rule to every elapsed task and returns
+	// the distinct jobs whose aggregate status may have changed.
+	ExpireLeases(ctx context.Context, now time.Time) ([]uuid.UUID, error)
 }
 
 // JobRepository persists jobs.
@@ -70,6 +79,9 @@ type JobRepository interface {
 	Insert(ctx context.Context, j *domain.Job) error
 	Get(ctx context.Context, id uuid.UUID) (*domain.Job, error)
 	UpdateStatus(ctx context.Context, id uuid.UUID, status domain.JobStatus, completedAt *time.Time) error
+	ClaimReduction(ctx context.Context, id uuid.UUID, startedAt time.Time) (bool, error)
+	CompleteWithResult(ctx context.Context, id, resultArtifactID uuid.UUID, completedAt time.Time) error
+	FailReduction(ctx context.Context, id uuid.UUID, code, message string, completedAt time.Time) error
 }
 
 // WorkerRepository persists the worker registry.
@@ -89,6 +101,9 @@ type WorkerRepository interface {
 type ArtifactRepository interface {
 	Insert(ctx context.Context, a *domain.Artifact) error
 	Get(ctx context.Context, id uuid.UUID) (*domain.Artifact, error)
+	// FindPartialResult returns the durable result already uploaded for one task
+	// attempt. A nil artifact means the attempt has not uploaded one yet.
+	FindPartialResult(ctx context.Context, taskID uuid.UUID, attempt int) (*domain.Artifact, error)
 }
 
 // BlobStore holds artifact bytes, addressed by an opaque storage key. It streams

coordinator/internal/usecase/preview.go

@@ -0,0 +1,151 @@
+package usecase
+
+import (
+	"context"
+	"encoding/csv"
+	"errors"
+	"io"
+	"mime"
+	"strings"
+
+	"github.com/google/uuid"
+
+	"github.com/emil28092005/SciMesh/coordinator/internal/domain"
+)
+
+// The preview is deliberately a diagnostic aid, never a full artifact
+// viewer. These limits bound both memory use and storage reads.
+const (
+	previewMaxRows  = 30
+	previewMaxBytes = 64 * 1024
+)
+
+// ArtifactPreviewView is the safe, bounded data rendered by the operator UI.
+// It deliberately contains neither storage keys nor worker-local details.
+type ArtifactPreviewView struct {
+	JobID       string
+	ArtifactID  string
+	Filename    string
+	Diagnostic  bool
+	Previewable bool
+	Reason      string
+	Headers     []string
+	Rows        [][]string
+	Truncated   bool
+	RowLimit    int
+	ByteLimit   int64
+}
+
+// PreviewArtifact reads the beginning of a job-scoped CSV result. Partial
+// results are diagnostic; a final result is available only after the reducer
+// has persisted it as this job's completed result.
+type PreviewArtifact struct {
+	read  UIReadRepository
+	blobs BlobStore
+}
+
+func NewPreviewArtifact(read UIReadRepository, blobs BlobStore) *PreviewArtifact {
+	return &PreviewArtifact{read: read, blobs: blobs}
+}
+
+func (p *PreviewArtifact) Execute(ctx context.Context, jobID, artifactID uuid.UUID) (ArtifactPreviewView, error) {
+	job, err := p.read.GetJob(ctx, jobID)
+	if err != nil {
+		return ArtifactPreviewView{}, err
+	}
+	// Another user's job (and not admin): report not-found, matching the
+	// artifact-absent response so nothing about it leaks.
+	if err := authorizeJobAccess(ctx, job); err != nil {
+		return ArtifactPreviewView{}, domain.ErrArtifactNotFound
+	}
+	artifacts, err := p.read.ListArtifactsByJob(ctx, jobID)
+	if err != nil {
+		return ArtifactPreviewView{}, err
+	}
+
+	var artifact *domain.Artifact
+	for i := range artifacts {
+		if artifacts[i].ID == artifactID {
+			artifact = &artifacts[i]
+			break
+		}
+	}
+	if artifact == nil || !previewableArtifact(*job, *artifact) {
+		// Use one response for an unknown artifact, another job's artifact, and
+		// an artifact that is not yet public. This avoids leaking its state.
+		return ArtifactPreviewView{}, domain.ErrArtifactNotFound
+	}
+
+	view := ArtifactPreviewView{
+		JobID:      jobID.String(),
+		ArtifactID: artifact.ID.String(),
+		Filename:   artifact.Filename,
+		Diagnostic: artifact.Kind == domain.ArtifactPartialResult,
+		RowLimit:   previewMaxRows,
+		ByteLimit:  previewMaxBytes,
+	}
+	if !isCSVArtifact(artifact) {
+		view.Reason = "This artifact is not a CSV file, so it cannot be shown as text here. Download it instead."
+		return view, nil
+	}
+	if artifact.SizeBytes == 0 {
+		view.Reason = "This artifact is empty."
+		return view, nil
+	}
+
+	body, err := p.blobs.Open(ctx, artifact.StorageKey)
+	if err != nil {
+		return ArtifactPreviewView{}, err
+	}
+	defer func() { _ = body.Close() }()
+
+	limited := &io.LimitedReader{R: body, N: previewMaxBytes}
+	reader := csv.NewReader(limited)
+	reader.FieldsPerRecord = -1 // a byte limit may end inside a record
+
+	headers, err := reader.Read()
+	if err != nil {
+		view.Reason = "This artifact could not be read as CSV."
+		return view, nil
+	}
+	view.Headers = append([]string(nil), headers...)
+	view.Rows = make([][]string, 0, previewMaxRows)
+	for len(view.Rows) < previewMaxRows {
+		record, readErr := reader.Read()
+		if readErr != nil {
+			if !errors.Is(readErr, io.EOF) {
+				view.Truncated = true
+			}
+			break
+		}
+		view.Rows = append(view.Rows, append([]string(nil), record...))
+	}
+
+	if artifact.SizeBytes > previewMaxBytes {
+		view.Truncated = true
+	} else if len(view.Rows) == previewMaxRows {
+		if _, readErr := reader.Read(); readErr == nil {
+			view.Truncated = true
+		}
+	}
+	view.Previewable = true
+	return view, nil
+}
+
+func previewableArtifact(job domain.Job, artifact domain.Artifact) bool {
+	if artifact.Kind == domain.ArtifactPartialResult {
+		return true
+	}
+	return artifact.Kind == domain.ArtifactFinalResult &&
+		job.Status == domain.JobCompleted &&
+		job.ResultArtifactID != nil &&
+		*job.ResultArtifactID == artifact.ID
+}
+
+func isCSVArtifact(artifact *domain.Artifact) bool {
+	mediaType, _, err := mime.ParseMediaType(artifact.ContentType)
+	if err == nil && strings.EqualFold(mediaType, "text/csv") {
+		return true
+	}
+	return strings.HasSuffix(strings.ToLower(artifact.Filename), ".csv")
+}

coordinator/internal/usecase/preview_test.go

@@ -0,0 +1,121 @@
+package usecase_test
+
+import (
+	"context"
+	"errors"
+	"strconv"
+	"strings"
+	"testing"
+	"time"
+
+	"github.com/google/uuid"
+
+	"github.com/emil28092005/SciMesh/coordinator/internal/domain"
+	"github.com/emil28092005/SciMesh/coordinator/internal/memstore"
+	"github.com/emil28092005/SciMesh/coordinator/internal/usecase"
+)
+
+func newPreviewHarness() (*usecase.PreviewArtifact, *memstore.JobRepo, *memstore.ArtifactRepo, *memstore.BlobStore) {
+	jobs := memstore.NewJobRepo()
+	tasks := memstore.NewTaskRepo()
+	workers := memstore.NewWorkerRepo()
+	artifacts := memstore.NewArtifactRepo()
+	blobs := memstore.NewBlobStore()
+	return usecase.NewPreviewArtifact(memstore.NewUIReadRepo(jobs, tasks, workers, artifacts), blobs), jobs, artifacts, blobs
+}
+
+func previewJob(t *testing.T, jobs *memstore.JobRepo, status domain.JobStatus) uuid.UUID {
+	t.Helper()
+	job := &domain.Job{ID: uuid.New(), Workload: "similarity-search", Status: status, CreatedAt: time.Now().UTC()}
+	if err := jobs.Insert(context.Background(), job); err != nil {
+		t.Fatalf("insert preview job: %v", err)
+	}
+	return job.ID
+}
+
+func previewArtifact(t *testing.T, artifacts *memstore.ArtifactRepo, blobs *memstore.BlobStore, jobID uuid.UUID, kind domain.ArtifactKind, filename, contentType, contents string) uuid.UUID {
+	t.Helper()
+	id := uuid.New()
+	sha, size, err := blobs.Put(context.Background(), id.String(), strings.NewReader(contents))
+	if err != nil {
+		t.Fatalf("store preview artifact: %v", err)
+	}
+	artifact := &domain.Artifact{ID: id, JobID: jobID, Kind: kind, Filename: filename, StorageKey: id.String(), ContentType: contentType, SizeBytes: size, SHA256: sha, CreatedAt: time.Now().UTC()}
+	if err := artifacts.Insert(context.Background(), artifact); err != nil {
+		t.Fatalf("insert preview artifact: %v", err)
+	}
+	return id
+}
+
+func TestPreviewArtifactRendersBoundedCSV(t *testing.T) {
+	preview, jobs, artifacts, blobs := newPreviewHarness()
+	jobID := previewJob(t, jobs, domain.JobRunning)
+	var csv strings.Builder
+	csv.WriteString("chembl_id,score\n")
+	for i := 0; i < 40; i++ {
+		csv.WriteString("CHEMBL" + strconv.Itoa(i) + ",0.9\n")
+	}
+	artifactID := previewArtifact(t, artifacts, blobs, jobID, domain.ArtifactPartialResult, "partial.csv", "text/csv; charset=utf-8", csv.String())
+
+	view, err := preview.Execute(context.Background(), jobID, artifactID)
+	if err != nil {
+		t.Fatalf("preview: %v", err)
+	}
+	if !view.Previewable || !view.Diagnostic || !view.Truncated || len(view.Rows) != 30 || view.Headers[0] != "chembl_id" || view.Rows[0][0] != "CHEMBL0" {
+		t.Fatalf("unexpected preview: %+v", view)
+	}
+}
+
+func TestPreviewArtifactCapsStorageReadAndHandlesInvalidCSV(t *testing.T) {
+	preview, jobs, artifacts, blobs := newPreviewHarness()
+	jobID := previewJob(t, jobs, domain.JobRunning)
+	// Fewer than 30 oversized records force the byte cap, rather than the row
+	// cap, to stop parsing.
+	large := "id,value\n" + strings.Repeat("row,"+strings.Repeat("x", 5*1024)+"\n", 20)
+	largeID := previewArtifact(t, artifacts, blobs, jobID, domain.ArtifactPartialResult, "large.csv", "text/csv", large)
+	view, err := preview.Execute(context.Background(), jobID, largeID)
+	if err != nil || !view.Previewable || !view.Truncated || len(view.Rows) > 30 {
+		t.Fatalf("large preview = (%+v, %v)", view, err)
+	}
+	invalidID := previewArtifact(t, artifacts, blobs, jobID, domain.ArtifactPartialResult, "broken.csv", "text/csv", "\"unterminated")
+	invalid, err := preview.Execute(context.Background(), jobID, invalidID)
+	if err != nil || invalid.Previewable || invalid.Reason == "" {
+		t.Fatalf("invalid preview = (%+v, %v)", invalid, err)
+	}
+}
+
+func TestPreviewArtifactRejectsOtherJobsAndNonResults(t *testing.T) {
+	preview, jobs, artifacts, blobs := newPreviewHarness()
+	jobA := previewJob(t, jobs, domain.JobRunning)
+	jobB := previewJob(t, jobs, domain.JobRunning)
+	partialID := previewArtifact(t, artifacts, blobs, jobA, domain.ArtifactPartialResult, "partial.csv", "text/csv", "a,b\n1,2\n")
+	if _, err := preview.Execute(context.Background(), jobB, partialID); !errors.Is(err, domain.ErrArtifactNotFound) {
+		t.Fatalf("cross-job preview error = %v", err)
+	}
+	inputID := previewArtifact(t, artifacts, blobs, jobA, domain.ArtifactInput, "input.csv", "text/csv", "a,b\n1,2\n")
+	if _, err := preview.Execute(context.Background(), jobA, inputID); !errors.Is(err, domain.ErrArtifactNotFound) {
+		t.Fatalf("input preview error = %v", err)
+	}
+}
+
+func TestPreviewArtifactExposesOnlyPersistedCompletedFinalResult(t *testing.T) {
+	preview, jobs, artifacts, blobs := newPreviewHarness()
+	jobID := previewJob(t, jobs, domain.JobReducing)
+	finalID := previewArtifact(t, artifacts, blobs, jobID, domain.ArtifactFinalResult, "final.csv", "text/csv", "rank,chembl_id\n1,CHEMBL1\n")
+	if _, err := preview.Execute(context.Background(), jobID, finalID); !errors.Is(err, domain.ErrArtifactNotFound) {
+		t.Fatalf("uncompleted final preview error = %v", err)
+	}
+	if err := jobs.CompleteWithResult(context.Background(), jobID, finalID, time.Now().UTC()); err != nil {
+		t.Fatal(err)
+	}
+	view, err := preview.Execute(context.Background(), jobID, finalID)
+	if err != nil || !view.Previewable || view.Diagnostic {
+		t.Fatalf("completed final preview = (%+v, %v)", view, err)
+	}
+	if err := jobs.FailReduction(context.Background(), jobID, "reducer_failed", "final result reduction failed", time.Now().UTC()); err != nil {
+		t.Fatal(err)
+	}
+	if _, err := preview.Execute(context.Background(), jobID, finalID); !errors.Is(err, domain.ErrArtifactNotFound) {
+		t.Fatalf("failed reducer preview error = %v", err)
+	}
+}

coordinator/internal/usecase/reduce.go

@@ -0,0 +1,136 @@
+package usecase
+
+import (
+	"bytes"
+	"context"
+	"io"
+
+	"github.com/google/uuid"
+
+	"github.com/emil28092005/SciMesh/coordinator/internal/domain"
+	"github.com/emil28092005/SciMesh/coordinator/internal/reducer"
+)
+
+// ReduceJob turns completed coordinator-owned partial artifacts into one final
+// artifact. It performs no worker I/O and never trusts a worker URI or path.
+type ReduceJob struct {
+	jobs      JobRepository
+	tasks     TaskRepository
+	artifacts ArtifactRepository
+	blobs     BlobStore
+	tx        TxManager
+	clock     Clock
+}
+
+func NewReduceJob(jobs JobRepository, tasks TaskRepository, artifacts ArtifactRepository,
+	blobs BlobStore, tx TxManager, clock Clock) *ReduceJob {
+	return &ReduceJob{jobs: jobs, tasks: tasks, artifacts: artifacts, blobs: blobs, tx: tx, clock: clock}
+}
+
+// Execute is idempotent for jobs that are not currently reducing. The worker
+// completion path may call it after every result; only the last task changes a
+// similarity-search job into reducing state.
+func (uc *ReduceJob) Execute(ctx context.Context, jobID uuid.UUID) error {
+	claimed, err := uc.jobs.ClaimReduction(ctx, jobID, uc.clock.Now())
+	if err != nil || !claimed {
+		return err
+	}
+	job, err := uc.jobs.Get(ctx, jobID)
+	if err != nil {
+		return err
+	}
+	if job.Status != domain.JobReducing {
+		return nil
+	}
+	if job.Workload != "similarity-search" {
+		return uc.fail(ctx, jobID)
+	}
+
+	completed, err := uc.tasks.ListCompleted(ctx, jobID)
+	if err != nil {
+		return uc.fail(ctx, jobID)
+	}
+	if len(completed) == 0 {
+		return uc.fail(ctx, jobID)
+	}
+	readers := make([]io.Reader, 0, len(completed))
+	closers := make([]io.Closer, 0, len(completed))
+	for _, task := range completed {
+		if task.ResultArtifactID == nil {
+			closeAll(closers)
+			return uc.fail(ctx, jobID)
+		}
+		artifact, err := uc.artifacts.Get(ctx, *task.ResultArtifactID)
+		if err != nil || artifact.JobID != jobID || artifact.TaskID == nil || *artifact.TaskID != task.ID || artifact.Kind != domain.ArtifactPartialResult {
+			closeAll(closers)
+			return uc.fail(ctx, jobID)
+		}
+		body, err := uc.blobs.Open(ctx, artifact.StorageKey)
+		if err != nil {
+			closeAll(closers)
+			return uc.fail(ctx, jobID)
+		}
+		readers = append(readers, body)
+		closers = append(closers, body)
+	}
+	output, reduceErr := reducer.ReduceSimilaritySearch(readers, job.Parameters)
+	closeAll(closers)
+	if reduceErr != nil {
+		return uc.fail(ctx, jobID)
+	}
+
+	final, err := domain.NewArtifact(jobID, nil, domain.ArtifactFinalResult, "similarity-search.csv", "text/csv", uc.clock.Now())
+	if err != nil {
+		return uc.fail(ctx, jobID)
+	}
+	sum, size, err := uc.blobs.Put(ctx, final.StorageKey, bytes.NewReader(output))
+	if err != nil {
+		return uc.fail(ctx, jobID)
+	}
+	final.SetContent(sum, size)
+	if err := uc.tx.WithinTx(ctx, func(ctx context.Context) error {
+		if err := uc.artifacts.Insert(ctx, final); err != nil {
+			return err
+		}
+		return uc.jobs.CompleteWithResult(ctx, jobID, final.ID, uc.clock.Now())
+	}); err != nil {
+		_ = uc.blobs.Delete(ctx, final.StorageKey)
+		return err
+	}
+	return nil
+}
+
+func (uc *ReduceJob) fail(ctx context.Context, jobID uuid.UUID) error {
+	// The public state carries a stable sanitized failure, never parser/storage
+	// internals that may include local paths or implementation details.
+	return uc.jobs.FailReduction(ctx, jobID, "reducer_failed", "final result reduction failed", uc.clock.Now())
+}
+
+func closeAll(closers []io.Closer) {
+	for _, closer := range closers {
+		_ = closer.Close()
+	}
+}
+
+type GetJobResult struct {
+	jobs     JobRepository
+	download *DownloadArtifact
+}
+
+func NewGetJobResult(jobs JobRepository, download *DownloadArtifact) *GetJobResult {
+	return &GetJobResult{jobs: jobs, download: download}
+}
+
+func (uc *GetJobResult) Execute(ctx context.Context, jobID uuid.UUID) (*domain.Artifact, io.ReadCloser, error) {
+	job, err := uc.jobs.Get(ctx, jobID)
+	if err != nil {
+		return nil, nil, err
+	}
+	if err := authorizeJobAccess(ctx, job); err != nil {
+		return nil, nil, err
+	}
+	if job.Status != domain.JobCompleted || job.ResultArtifactID == nil {
+		return nil, nil, domain.ErrArtifactNotFound
+	}
+	return uc.download.Execute(ctx, *job.ResultArtifactID)
+}

coordinator/internal/usecase/task.go

@@ -2,6 +2,7 @@ package usecase
 
 import (
 	"context"
+	"errors"
 	"time"
 
 	"github.com/google/uuid"
@@ -21,12 +22,15 @@ import (
 
 type ClaimTask struct {
 	tasks         TaskRepository
+	jobs          JobRepository
+	workers       WorkerRepository
+	tx            TxManager
 	clock         Clock
 	leaseDuration time.Duration
 }
 
-func NewClaimTask(tasks TaskRepository, clock Clock, leaseDuration time.Duration) *ClaimTask {
-	return &ClaimTask{tasks: tasks, clock: clock, leaseDuration: leaseDuration}
+func NewClaimTask(tasks TaskRepository, jobs JobRepository, workers WorkerRepository, tx TxManager, clock Clock, leaseDuration time.Duration) *ClaimTask {
+	return &ClaimTask{tasks: tasks, jobs: jobs, workers: workers, tx: tx, clock: clock, leaseDuration: leaseDuration}
 }
 
 // Execute reclaims elapsed leases first, then hands out one task.
@@ -42,27 +46,61 @@ func (uc *ClaimTask) Execute(ctx context.Context, in ClaimTaskInput) (*domain.Cl
 	if in.WorkerID == "" {
 		return nil, domain.ErrInvalidInput
 	}
-	now := uc.clock.Now()
-
-	if _, err := uc.tasks.ExpireLeases(ctx, now); err != nil {
+	workloads := in.Workloads
+	var voterOwner *uuid.UUID
+	if workerID, err := uuid.Parse(in.WorkerID); err == nil {
+		worker, err := uc.workers.Get(ctx, workerID)
+		if err != nil {
 			return nil, err
 		}
+		// Bind the caller to the worker it claims as. A JWT-authenticated
+		// volunteer may operate only its own workers; without this the trust
+		// tier would be read off a caller-supplied worker_id, letting anyone who
+		// knows a trusted worker's id claim as it. A shared-token caller (no
+		// requester) is a lab operator and may act as any worker.
+		if err := authorizeWorkerOwner(ctx, uc.workers, in.WorkerID); err != nil {
+			return nil, err
+		}
+		// An untrusted volunteer may claim, but never a chunk its owner has
+		// already voted on — so quorum needs genuinely independent computations.
+		if worker.TrustLevel == domain.WorkerUntrusted {
+			voterOwner = worker.OwnerID
+		}
+		// Never trust caller-supplied capabilities: registration is the durable
+		// worker identity and its allowlist.
+		workloads = worker.Capabilities
+	}
+	var claimed *domain.ClaimedTask
+	err := uc.tx.WithinTx(ctx, func(ctx context.Context) error {
+		now := uc.clock.Now()
+		affectedJobs, err := uc.tasks.ExpireLeases(ctx, now)
+		if err != nil {
+			return err
+		}
+		if err := syncExpiredJobStatuses(ctx, uc.jobs, uc.tasks, affectedJobs, now); err != nil {
+			return err
+		}
 
 		task, err := uc.tasks.ClaimNext(ctx, ClaimFilter{
-		Workloads:  in.Workloads,
+			Workloads:  workloads,
 			Owner:      in.WorkerID,
 			Now:        now,
 			LeaseUntil: now.Add(uc.leaseDuration),
+			VoterOwner: voterOwner,
+		})
+		if err != nil {
+			return err
+		}
+		if task != nil {
+			value := task.AsClaimed()
+			claimed = &value
+		}
+		return nil
 	})
 	if err != nil {
 		return nil, err
 	}
-	if task == nil {
-		return nil, nil // empty queue is a normal state, not an error
-	}
-
-	claimed := task.AsClaimed()
-	return &claimed, nil
+	return claimed, nil // nil means an empty queue
 }
 
 // --- RenewLease ----------------------------------------------------------
@@ -84,6 +122,9 @@ func NewRenewLease(tasks TaskRepository, workers WorkerRepository, tx TxManager,
 // locked: two concurrent heartbeats must not interleave into a lost update.
 // Whether the caller may renew at all is decided by the entity, not here.
 func (uc *RenewLease) Execute(ctx context.Context, in RenewLeaseInput) (*domain.ClaimedTask, error) {
+	if err := authorizeWorkerOwner(ctx, uc.workers, in.WorkerID); err != nil {
+		return nil, err
+	}
 	var claimed domain.ClaimedTask
 
 	err := uc.tx.WithinTx(ctx, func(ctx context.Context) error {
@@ -119,13 +160,22 @@ type CompleteTask struct {
 	tasks     TaskRepository
 	jobs      JobRepository
 	artifacts ArtifactRepository
+	workers   WorkerRepository
+	results   TaskResultRepository
 	tx        TxManager
 	clock     Clock
+	// quorum is how many distinct owners must agree on an untrusted result
+	// before it is accepted; a trusted worker's result is accepted directly.
+	quorum int
 }
 
 func NewCompleteTask(tasks TaskRepository, jobs JobRepository, artifacts ArtifactRepository,
-	tx TxManager, clock Clock) *CompleteTask {
-	return &CompleteTask{tasks: tasks, jobs: jobs, artifacts: artifacts, tx: tx, clock: clock}
+	workers WorkerRepository, results TaskResultRepository, tx TxManager, clock Clock, quorum int) *CompleteTask {
+	if quorum < 1 {
+		quorum = 2
+	}
+	return &CompleteTask{tasks: tasks, jobs: jobs, artifacts: artifacts, workers: workers,
+		results: results, tx: tx, clock: clock, quorum: quorum}
 }
 
 // Execute applies the result and, when that was the job's last outstanding
@@ -135,6 +185,9 @@ func NewCompleteTask(tasks TaskRepository, jobs JobRepository, artifacts Artifac
 // Lease ownership, staleness, and idempotent replays are all decided by
 // Task.CompleteWith; this use case only orchestrates.
 func (uc *CompleteTask) Execute(ctx context.Context, in CompleteTaskInput) (*domain.Task, error) {
+	if err := authorizeWorkerOwner(ctx, uc.workers, in.WorkerID); err != nil {
+		return nil, err
+	}
 	var out *domain.Task
 
 	err := uc.tx.WithinTx(ctx, func(ctx context.Context) error {
@@ -144,24 +197,35 @@ func (uc *CompleteTask) Execute(ctx context.Context, in CompleteTaskInput) (*dom
 		}
 		// Rule 10: never trust a worker-supplied artifact reference. The result
 		// must be an artifact the coordinator itself stored for *this* task.
-		if err := uc.verifyResultArtifact(ctx, in.TaskID, in.Attempt, in.ResultArtifactID); err != nil {
+		art, err := uc.verifyResultArtifact(ctx, in.TaskID, in.Attempt, in.ResultArtifactID)
+		if err != nil {
+			return err
+		}
+
+		trusted, ownerID, err := uc.workerTrust(ctx, in.WorkerID)
+		if err != nil {
 			return err
 		}
 		now := uc.clock.Now()
+
+		// Untrusted (volunteer) worker: record a vote and only complete once a
+		// quorum of distinct owners agree; otherwise return the task to the queue.
+		if !trusted {
+			return uc.recordVote(ctx, task, in, art, ownerID, now, &out)
+		}
+
+		// Trusted worker (lab token, verified, or admin): accept directly.
 		before := task.Version
-		if err := task.CompleteWith(in.ResultArtifactID, in.Metrics,
-			in.WorkerID, in.Attempt, now); err != nil {
+		if err := task.CompleteWith(in.ResultArtifactID, in.Metrics, in.WorkerID, in.Attempt, now); err != nil {
 			return err
 		}
 		out = task
-
 		// A replay of an already-recorded result leaves the entity untouched.
 		// Writing anyway would fail the optimistic-concurrency guard (the stored
 		// version already equals ours) and turn an idempotent call into a 409.
 		if task.Version == before {
 			return nil
 		}
-
 		if err := uc.tasks.Update(ctx, task); err != nil {
 			return err
 		}
@@ -173,18 +237,83 @@ func (uc *CompleteTask) Execute(ctx context.Context, in CompleteTaskInput) (*dom
 	return out, nil
 }
 
-// verifyResultArtifact enforces that the referenced artifact was stored by the
-// coordinator for this exact task. It stops a worker from completing task B with
-// an artifact it uploaded for task A, and from naming an id that isn't a result.
-func (uc *CompleteTask) verifyResultArtifact(ctx context.Context, taskID uuid.UUID, attempt int, artifactID uuid.UUID) error {
-	art, err := uc.artifacts.Get(ctx, artifactID)
+// recordVote handles an untrusted result: it stores the vote, then completes the
+// task when the submitter's result hash has reached quorum, or returns the task
+// to the queue so another owner can compute it independently.
+func (uc *CompleteTask) recordVote(ctx context.Context, task *domain.Task, in CompleteTaskInput,
+	art *domain.Artifact, ownerID uuid.UUID, now time.Time, out **domain.Task) error {
+
+	*out = task
+	if task.Status == domain.TaskCompleted {
+		return nil // already settled by an earlier quorum; nothing to record
+	}
+	if err := uc.results.RecordVote(ctx, task.ID, ownerID, art.SHA256, in.ResultArtifactID); err != nil {
+		return err
+	}
+	agree, err := uc.results.CountAgreeing(ctx, task.ID, art.SHA256)
 	if err != nil {
 		return err
 	}
-	if art.TaskID == nil || *art.TaskID != taskID || art.Attempt == nil || *art.Attempt != attempt || art.Kind != domain.ArtifactPartialResult {
-		return domain.ErrResultConflict
+
+	if agree >= uc.quorum {
+		// The submitter's own (already verified) artifact carries the winning
+		// hash, so complete with it.
+		if err := task.CompleteWith(in.ResultArtifactID, in.Metrics, in.WorkerID, in.Attempt, now); err != nil {
+			return err
 		}
-	return nil
+	} else if err := task.ReleaseAfterVote(in.WorkerID, in.Attempt, now); err != nil {
+		return err
+	}
+
+	if err := uc.tasks.Update(ctx, task); err != nil {
+		return err
+	}
+	return syncJobStatus(ctx, uc.jobs, uc.tasks, task.JobID, now)
+}
+
+// workerTrust reports whether the worker's results are accepted directly, and
+// the owner to attribute a vote to when they are not.
+func (uc *CompleteTask) workerTrust(ctx context.Context, workerID string) (trusted bool, ownerID uuid.UUID, err error) {
+	// When the worker can't be resolved, default to trusted — the pre-quorum
+	// behaviour. This is safe because completing a task requires holding its
+	// lease, and the lease owner is always a real registered worker whose trust
+	// is therefore known; only an untrusted worker ever takes the quorum path.
+	id, err := uuid.Parse(workerID)
+	if err != nil {
+		// An unparseable worker id means the worker can't be resolved; fall back
+		// to the trusted default rather than surfacing the parse error.
+		return true, uuid.Nil, nil //nolint:nilerr // unresolvable worker → trusted (pre-quorum default)
+	}
+	w, err := uc.workers.Get(ctx, id)
+	if err != nil {
+		if errors.Is(err, domain.ErrWorkerNotFound) {
+			return true, uuid.Nil, nil
+		}
+		return false, uuid.Nil, err
+	}
+	if w.TrustLevel != domain.WorkerUntrusted {
+		return true, uuid.Nil, nil
+	}
+	if w.OwnerID == nil {
+		// An untrusted worker always has an owner (it registered via a user JWT);
+		// a missing one is a data error, not a silent trust upgrade.
+		return false, uuid.Nil, domain.ErrInvalidInput
+	}
+	return false, *w.OwnerID, nil
+}
+
+// verifyResultArtifact enforces that the referenced artifact was stored by the
+// coordinator for this exact task. It stops a worker from completing task B with
+// an artifact it uploaded for task A, and from naming an id that isn't a result.
+func (uc *CompleteTask) verifyResultArtifact(ctx context.Context, taskID uuid.UUID, attempt int, artifactID uuid.UUID) (*domain.Artifact, error) {
+	art, err := uc.artifacts.Get(ctx, artifactID)
+	if err != nil {
+		return nil, err
+	}
+	if art.TaskID == nil || *art.TaskID != taskID || art.Attempt == nil || *art.Attempt != attempt || art.Kind != domain.ArtifactPartialResult {
+		return nil, domain.ErrResultConflict
+	}
+	return art, nil
 }
 
 // --- FailTask ------------------------------------------------------------
@@ -192,17 +321,21 @@ func (uc *CompleteTask) verifyResultArtifact(ctx context.Context, taskID uuid.UU
 type FailTask struct {
 	tasks   TaskRepository
 	jobs    JobRepository
+	workers WorkerRepository
 	tx      TxManager
 	clock   Clock
 }
 
-func NewFailTask(tasks TaskRepository, jobs JobRepository, tx TxManager, clock Clock) *FailTask {
-	return &FailTask{tasks: tasks, jobs: jobs, tx: tx, clock: clock}
+func NewFailTask(tasks TaskRepository, jobs JobRepository, workers WorkerRepository, tx TxManager, clock Clock) *FailTask {
+	return &FailTask{tasks: tasks, jobs: jobs, workers: workers, tx: tx, clock: clock}
 }
 
 // Execute delegates the requeue-or-terminate decision to Task.Fail, then keeps
 // the parent job's status consistent in the same transaction.
 func (uc *FailTask) Execute(ctx context.Context, in FailTaskInput) (*domain.Task, error) {
+	if err := authorizeWorkerOwner(ctx, uc.workers, in.WorkerID); err != nil {
+		return nil, err
+	}
 	var out *domain.Task
 
 	err := uc.tx.WithinTx(ctx, func(ctx context.Context) error {
@@ -230,18 +363,30 @@ func (uc *FailTask) Execute(ctx context.Context, in FailTaskInput) (*domain.Task
 
 type ExpireLeases struct {
 	tasks TaskRepository
+	jobs  JobRepository
+	tx    TxManager
 	clock Clock
 }
 
-func NewExpireLeases(tasks TaskRepository, clock Clock) *ExpireLeases {
-	return &ExpireLeases{tasks: tasks, clock: clock}
+func NewExpireLeases(tasks TaskRepository, jobs JobRepository, tx TxManager, clock Clock) *ExpireLeases {
+	return &ExpireLeases{tasks: tasks, jobs: jobs, tx: tx, clock: clock}
 }
 
-// Execute reports how many tasks were reclaimed.
+// Execute reclaims elapsed tasks and persists the state of every affected job.
 //
 // The sweep is one set-based statement rather than a load-decide-save loop:
 // several coordinators run it concurrently, and a single atomic UPDATE makes
 // the duplicate work harmless — the loser simply updates 0 rows.
 func (uc *ExpireLeases) Execute(ctx context.Context) (int64, error) {
-	return uc.tasks.ExpireLeases(ctx, uc.clock.Now())
+	var affected []uuid.UUID
+	err := uc.tx.WithinTx(ctx, func(ctx context.Context) error {
+		now := uc.clock.Now()
+		var err error
+		affected, err = uc.tasks.ExpireLeases(ctx, now)
+		if err != nil {
+			return err
+		}
+		return syncExpiredJobStatuses(ctx, uc.jobs, uc.tasks, affected, now)
+	})
+	return int64(len(affected)), err
 }

coordinator/internal/usecase/ui.go

@@ -2,10 +2,12 @@ package usecase
 
 import (
 	"context"
+	"fmt"
 	"time"
 
 	"github.com/google/uuid"
 
+	"github.com/emil28092005/SciMesh/coordinator/internal/authctx"
 	"github.com/emil28092005/SciMesh/coordinator/internal/domain"
 )
 
@@ -13,8 +15,11 @@ import (
 // It intentionally exposes no storage paths or credentials.
 type UIReadRepository interface {
 	GetJob(ctx context.Context, jobID uuid.UUID) (*domain.Job, error)
-	ListJobs(ctx context.Context, limit int) ([]domain.Job, error)
+	// ListJobs returns the most recent jobs. A non-nil owner restricts the list
+	// to that user's jobs; nil returns all (operator/admin view).
+	ListJobs(ctx context.Context, owner *uuid.UUID, limit int) ([]domain.Job, error)
 	ListTasksByJob(ctx context.Context, jobID uuid.UUID) ([]domain.Task, error)
+	ListTasksByJobs(ctx context.Context, jobIDs []uuid.UUID) (map[uuid.UUID][]domain.Task, error)
 	ListWorkers(ctx context.Context, limit int) ([]domain.Worker, error)
 	ListArtifactsByJob(ctx context.Context, jobID uuid.UUID) ([]domain.Artifact, error)
 }
@@ -24,6 +29,10 @@ type JobCard struct {
 	Workload         string     `json:"workload"`
 	Status           string     `json:"status"`
 	CreatedAt        time.Time  `json:"created_at"`
+	CompletedAt      *time.Time `json:"completed_at,omitempty"`
+	ReducerStartedAt *time.Time `json:"reducer_started_at,omitempty"`
+	ErrorCode        string     `json:"error_code,omitempty"`
+	ErrorMessage     string     `json:"error_message,omitempty"`
 	Total            int        `json:"total"`
 	Pending          int        `json:"pending"`
 	Leased           int        `json:"leased"`
@@ -41,10 +50,20 @@ type TaskCard struct {
 	MaxAttempts    int        `json:"max_attempts"`
 	LeaseOwner     string     `json:"lease_owner,omitempty"`
 	LeaseExpiresAt *time.Time `json:"lease_expires_at,omitempty"`
+	StartedAt      *time.Time `json:"started_at,omitempty"`
+	CompletedAt    *time.Time `json:"completed_at,omitempty"`
 	ErrorCode      string     `json:"error_code,omitempty"`
 	ErrorMessage   string     `json:"error_message,omitempty"`
 }
 
+// ParameterCard is an intentionally small allowlist of run configuration that
+// helps an operator verify what is being computed without exposing arbitrary
+// job payloads to the browser.
+type ParameterCard struct {
+	Label string `json:"label"`
+	Value string `json:"value"`
+}
+
 type ArtifactCard struct {
 	ID           string `json:"id"`
 	Kind         string `json:"kind"`
@@ -64,14 +83,40 @@ type WorkerCard struct {
 }
 
 type DashboardView struct {
-	Jobs    []JobCard
-	Workers []WorkerCard
+	Jobs          []JobCard    `json:"jobs"`
+	Workers       []WorkerCard `json:"workers"`
+	ActiveJobs    int          `json:"active_jobs"`
+	FinishedJobs  int          `json:"finished_jobs"`
+	OnlineWorkers int          `json:"online_workers"`
+	// Session is the signed-in user, when the UI runs in session mode. nil under
+	// basic auth. Template-only, never serialised to the polling JSON.
+	Session *SessionView `json:"-"`
 }
+
+// SessionView is the minimal identity the UI header needs to show who is signed
+// in and to offer a logout control.
+type SessionView struct {
+	Role     string
+	Verified bool
+}
+
+// sessionViewFrom builds the header session info from the request context, or
+// nil when the caller is not an authenticated user (basic-auth operator).
+func sessionViewFrom(ctx context.Context) *SessionView {
+	r, ok := authctx.From(ctx)
+	if !ok {
+		return nil
+	}
+	return &SessionView{Role: r.Role, Verified: r.Verified}
+}
+
 type JobDetailView struct {
 	JobCard
 	Tasks                []TaskCard      `json:"tasks"`
 	Artifacts            []ArtifactCard  `json:"artifacts"`
+	Parameters           []ParameterCard `json:"parameters"`
 	FinalResultAvailable bool            `json:"final_result_available"`
+	Session              *SessionView    `json:"-"`
 }
 
 type Dashboard struct{ read UIReadRepository }
@@ -79,7 +124,7 @@ type Dashboard struct{ read UIReadRepository }
 func NewDashboard(read UIReadRepository) *Dashboard { return &Dashboard{read: read} }
 
 func (d *Dashboard) Overview(ctx context.Context, limit int) (DashboardView, error) {
-	jobs, err := d.read.ListJobs(ctx, limit)
+	jobs, err := d.read.ListJobs(ctx, uiOwnerFilter(ctx), limit)
 	if err != nil {
 		return DashboardView{}, err
 	}
@@ -88,16 +133,31 @@ func (d *Dashboard) Overview(ctx context.Context, limit int) (DashboardView, err
 		return DashboardView{}, err
 	}
 	out := DashboardView{Jobs: make([]JobCard, 0, len(jobs)), Workers: make([]WorkerCard, 0, len(workers))}
+	jobIDs := make([]uuid.UUID, 0, len(jobs))
 	for _, job := range jobs {
-		tasks, err := d.read.ListTasksByJob(ctx, job.ID)
+		jobIDs = append(jobIDs, job.ID)
+	}
+	tasksByJob, err := d.read.ListTasksByJobs(ctx, jobIDs)
 	if err != nil {
 		return DashboardView{}, err
 	}
-		out.Jobs = append(out.Jobs, jobCard(job, tasks))
+	for _, job := range jobs {
+		card := jobCard(job, tasksByJob[job.ID])
+		out.Jobs = append(out.Jobs, card)
+		switch card.Status {
+		case string(domain.JobCompleted), string(domain.JobFailed), string(domain.JobCancelled):
+			out.FinishedJobs++
+		default:
+			out.ActiveJobs++
+		}
 	}
 	for _, worker := range workers {
 		out.Workers = append(out.Workers, WorkerCard{ID: worker.ID.String(), Name: worker.Name, Status: string(worker.Status), Capabilities: worker.Capabilities, LastHeartbeatAt: worker.LastHeartbeatAt})
+		if worker.Status == domain.WorkerOnline || worker.Status == domain.WorkerBusy {
+			out.OnlineWorkers++
 		}
+	}
+	out.Session = sessionViewFrom(ctx)
 	return out, nil
 }
 
@@ -106,6 +166,11 @@ func (d *Dashboard) JobDetail(ctx context.Context, jobID uuid.UUID) (JobDetailVi
 	if err != nil {
 		return JobDetailView{}, err
 	}
+	// A plain user may only open their own job; a mismatch reads as not-found so
+	// the page never reveals another user's job exists.
+	if err := authorizeJobAccess(ctx, job); err != nil {
+		return JobDetailView{}, err
+	}
 	tasks, err := d.read.ListTasksByJob(ctx, jobID)
 	if err != nil {
 		return JobDetailView{}, err
@@ -114,11 +179,28 @@ func (d *Dashboard) JobDetail(ctx context.Context, jobID uuid.UUID) (JobDetailVi
 	if err != nil {
 		return JobDetailView{}, err
 	}
-	out := JobDetailView{JobCard: jobCard(*job, tasks), Tasks: make([]TaskCard, 0, len(tasks)), Artifacts: make([]ArtifactCard, 0, len(artifacts))}
+	workers, err := d.read.ListWorkers(ctx, 100)
+	if err != nil {
+		return JobDetailView{}, err
+	}
+	workerNames := make(map[string]string, len(workers))
+	for _, worker := range workers {
+		workerNames[worker.ID.String()] = worker.Name
+	}
+	out := JobDetailView{
+		JobCard:    jobCard(*job, tasks),
+		Tasks:      make([]TaskCard, 0, len(tasks)),
+		Artifacts:  make([]ArtifactCard, 0, len(artifacts)),
+		Parameters: uiParameters(job.Parameters),
+		Session:    sessionViewFrom(ctx),
+	}
 	for _, task := range tasks {
-		card := TaskCard{ID: task.ID.String(), ChunkIndex: task.ChunkIndex, Status: string(task.Status), Attempt: task.Attempt, MaxAttempts: task.MaxAttempts, LeaseExpiresAt: task.LeaseExpiresAt}
+		card := TaskCard{ID: task.ID.String(), ChunkIndex: task.ChunkIndex, Status: string(task.Status), Attempt: task.Attempt, MaxAttempts: task.MaxAttempts, LeaseExpiresAt: task.LeaseExpiresAt, StartedAt: task.StartedAt, CompletedAt: task.CompletedAt}
 		if task.LeaseOwner != nil {
-			card.LeaseOwner = *task.LeaseOwner
+			card.LeaseOwner = workerNames[*task.LeaseOwner]
+			if card.LeaseOwner == "" {
+				card.LeaseOwner = "Worker " + shortID(*task.LeaseOwner)
+			}
 		}
 		if task.ErrorCode != nil {
 			card.ErrorCode = *task.ErrorCode
@@ -130,7 +212,7 @@ func (d *Dashboard) JobDetail(ctx context.Context, jobID uuid.UUID) (JobDetailVi
 	}
 	for _, artifact := range artifacts {
 		diagnostic := artifact.Kind == domain.ArtifactPartialResult
-		downloadable := diagnostic || (artifact.Kind == domain.ArtifactFinalResult && out.Status == string(domain.JobCompleted))
+		downloadable := previewableArtifact(*job, artifact)
 		out.Artifacts = append(out.Artifacts, ArtifactCard{ID: artifact.ID.String(), Kind: string(artifact.Kind), Filename: artifact.Filename, SizeBytes: artifact.SizeBytes, SHA256: artifact.SHA256, Downloadable: downloadable, Diagnostic: diagnostic})
 		if artifact.Kind == domain.ArtifactFinalResult && downloadable {
 			out.FinalResultAvailable = true
@@ -139,13 +221,25 @@ func (d *Dashboard) JobDetail(ctx context.Context, jobID uuid.UUID) (JobDetailVi
 	return out, nil
 }
 
-func (d *Dashboard) ArtifactBelongsToJob(ctx context.Context, jobID, artifactID uuid.UUID) (bool, error) {
+// DownloadableArtifactBelongsToJob applies the same policy used by the UI
+// projection: partial diagnostics and the persisted final result are public to
+// the operator; source inputs and shards are not exposed through a guessed UI
+// URL.
+func (d *Dashboard) DownloadableArtifactBelongsToJob(ctx context.Context, jobID, artifactID uuid.UUID) (bool, error) {
+	job, err := d.read.GetJob(ctx, jobID)
+	if err != nil {
+		return false, err
+	}
+	// Not the caller's job (and not admin): treat as if the artifact is absent.
+	if err := authorizeJobAccess(ctx, job); err != nil {
+		return false, nil //nolint:nilerr // masking the authz error as "not found" is intentional
+	}
 	artifacts, err := d.read.ListArtifactsByJob(ctx, jobID)
 	if err != nil {
 		return false, err
 	}
 	for _, a := range artifacts {
-		if a.ID == artifactID {
+		if a.ID == artifactID && previewableArtifact(*job, a) {
 			return true, nil
 		}
 	}
@@ -153,7 +247,13 @@ func (d *Dashboard) ArtifactBelongsToJob(ctx context.Context, jobID, artifactID
 }
 
 func jobCard(job domain.Job, tasks []domain.Task) JobCard {
-	c := JobCard{ID: job.ID.String(), Workload: job.Workload, CreatedAt: job.CreatedAt}
+	c := JobCard{ID: job.ID.String(), Workload: job.Workload, CreatedAt: job.CreatedAt, CompletedAt: job.CompletedAt, ReducerStartedAt: job.ReducerStartedAt}
+	if job.ErrorCode != nil {
+		c.ErrorCode = *job.ErrorCode
+	}
+	if job.ErrorMessage != nil {
+		c.ErrorMessage = *job.ErrorMessage
+	}
 	for _, task := range tasks {
 		c.Total++
 		switch task.Status {
@@ -175,3 +275,46 @@ func jobCard(job domain.Job, tasks []domain.Task) JobCard {
 	c.Status = string(p.DeriveStatus())
 	return c
 }
+
+func uiParameters(parameters map[string]any) []ParameterCard {
+	keys := []struct {
+		key   string
+		label string
+	}{
+		{"query_smiles", "Target SMILES"},
+		{"query_id", "Target ChEMBL ID"},
+		{"top_k", "Global top-k"},
+		{"threshold", "Similarity threshold"},
+		{"threshold_direction", "Threshold direction"},
+	}
+	out := make([]ParameterCard, 0, len(keys))
+	for _, entry := range keys {
+		value, ok := parameters[entry.key]
+		if !ok {
+			continue
+		}
+		formatted, ok := formatUIParameter(value)
+		if ok {
+			out = append(out, ParameterCard{Label: entry.label, Value: formatted})
+		}
+	}
+	return out
+}
+
+func formatUIParameter(value any) (string, bool) {
+	switch typed := value.(type) {
+	case string:
+		return typed, true
+	case int, int64, float64, bool:
+		return fmt.Sprint(typed), true
+	default:
+		return "", false
+	}
+}
+
+func shortID(value string) string {
+	if len(value) <= 8 {
+		return value
+	}
+	return value[:8]
+}

coordinator/internal/usecase/ui_internal_test.go

@@ -0,0 +1,19 @@
+package usecase
+
+import "testing"
+
+func TestUIParametersAreAllowlisted(t *testing.T) {
+	parameters := uiParameters(map[string]any{
+		"query_smiles":         "CCO",
+		"top_k":                float64(20),
+		"internal_storage_key": "must-not-reach-browser",
+		"nested":               map[string]any{"secret": "no"},
+	})
+	if len(parameters) != 2 {
+		t.Fatalf("parameters = %#v, want only two allowlisted values", parameters)
+	}
+	if parameters[0] != (ParameterCard{Label: "Target SMILES", Value: "CCO"}) ||
+		parameters[1] != (ParameterCard{Label: "Global top-k", Value: "20"}) {
+		t.Fatalf("parameters = %#v", parameters)
+	}
+}

coordinator/internal/usecase/ui_scope_test.go

@@ -0,0 +1,82 @@
+package usecase_test
+
+import (
+	"context"
+	"errors"
+	"testing"
+	"time"
+
+	"github.com/google/uuid"
+
+	"github.com/emil28092005/SciMesh/coordinator/internal/authctx"
+	"github.com/emil28092005/SciMesh/coordinator/internal/domain"
+	"github.com/emil28092005/SciMesh/coordinator/internal/memstore"
+	"github.com/emil28092005/SciMesh/coordinator/internal/usecase"
+)
+
+func newDashboard() (*usecase.Dashboard, *memstore.JobRepo) {
+	jobs := memstore.NewJobRepo()
+	tasks := memstore.NewTaskRepo()
+	workers := memstore.NewWorkerRepo()
+	artifacts := memstore.NewArtifactRepo()
+	return usecase.NewDashboard(memstore.NewUIReadRepo(jobs, tasks, workers, artifacts)), jobs
+}
+
+func ownedJob(t *testing.T, jobs *memstore.JobRepo, owner uuid.UUID) uuid.UUID {
+	t.Helper()
+	o := owner
+	job := &domain.Job{ID: uuid.New(), Workload: "similarity-search", Status: domain.JobRunning, OwnerID: &o, CreatedAt: time.Now().UTC()}
+	if err := jobs.Insert(context.Background(), job); err != nil {
+		t.Fatalf("insert owned job: %v", err)
+	}
+	return job.ID
+}
+
+func userCtx(id uuid.UUID, role string) context.Context {
+	return authctx.With(context.Background(), authctx.Requester{UserID: id, Role: role})
+}
+
+func TestOverviewScopesJobsByOwner(t *testing.T) {
+	dash, jobs := newDashboard()
+	alice, bob := uuid.New(), uuid.New()
+	ownedJob(t, jobs, alice)
+	ownedJob(t, jobs, bob)
+
+	// A plain user sees only their own job.
+	v, err := dash.Overview(userCtx(alice, "user"), 20)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if len(v.Jobs) != 1 {
+		t.Errorf("alice sees %d jobs, want 1", len(v.Jobs))
+	}
+
+	// An admin sees every job.
+	if v, _ := dash.Overview(userCtx(uuid.New(), "admin"), 20); len(v.Jobs) != 2 {
+		t.Errorf("admin sees %d jobs, want 2", len(v.Jobs))
+	}
+
+	// No requester (basic-auth operator) sees every job — unchanged behaviour.
+	if v, _ := dash.Overview(context.Background(), 20); len(v.Jobs) != 2 {
+		t.Errorf("operator sees %d jobs, want 2", len(v.Jobs))
+	}
+}
+
+func TestJobDetailRejectsAnotherUsersJob(t *testing.T) {
+	dash, jobs := newDashboard()
+	alice, bob := uuid.New(), uuid.New()
+	jobID := ownedJob(t, jobs, alice)
+
+	// Bob cannot open Alice's job.
+	if _, err := dash.JobDetail(userCtx(bob, "user"), jobID); !errors.Is(err, domain.ErrJobNotFound) {
+		t.Errorf("bob: got %v, want ErrJobNotFound", err)
+	}
+	// Alice can.
+	if _, err := dash.JobDetail(userCtx(alice, "user"), jobID); err != nil {
+		t.Errorf("alice: unexpected error %v", err)
+	}
+	// Admin can.
+	if _, err := dash.JobDetail(userCtx(uuid.New(), "admin"), jobID); err != nil {
+		t.Errorf("admin: unexpected error %v", err)
+	}
+}

coordinator/internal/usecase/upload.go

@@ -4,6 +4,7 @@ import (
 	"context"
 	"fmt"
 	"io"
+	"math"
 
 	"github.com/google/uuid"
 
@@ -21,20 +22,28 @@ type SubmitDataset struct {
 	tasks       TaskRepository
 	tx          TxManager
 	clk         Clock
+	maxAttempts int
 }
 
 func NewSubmitDataset(blobs BlobStore, artifacts ArtifactRepository, jobs JobRepository,
-	tasks TaskRepository, tx TxManager, clk Clock) *SubmitDataset {
-	return &SubmitDataset{blobs: blobs, artifacts: artifacts, jobs: jobs, tasks: tasks, tx: tx, clk: clk}
+	tasks TaskRepository, tx TxManager, clk Clock, maxAttempts int) *SubmitDataset {
+	return &SubmitDataset{blobs: blobs, artifacts: artifacts, jobs: jobs, tasks: tasks, tx: tx, clk: clk, maxAttempts: maxAttempts}
 }
 
 func (uc *SubmitDataset) Execute(ctx context.Context, in SubmitDatasetInput) (SubmitDatasetResult, error) {
+	if err := validateUploadedWorkload(in.Workload, in.Parameters); err != nil {
+		return SubmitDatasetResult{}, err
+	}
+	if uc.maxAttempts < 1 {
+		return SubmitDatasetResult{}, domain.ErrInvalidInput
+	}
 	now := uc.clk.Now()
 
 	job, err := domain.NewUploadedJob(in.Workload, in.Parameters, now)
 	if err != nil {
 		return SubmitDatasetResult{}, err
 	}
+	job.OwnerID = ownerFromContext(ctx)
 
 	// Everything written to blob storage, so a failed transaction can undo it.
 	var putKeys []string
@@ -64,7 +73,7 @@ func (uc *SubmitDataset) Execute(ctx context.Context, in SubmitDatasetInput) (Su
 		cleanup()
 		return SubmitDatasetResult{}, err
 	}
-	splitErr := chunk.SplitTSVLimit(rc, in.RowsPerShard, in.MaxRows, func(index int, shard io.Reader) error {
+	splitErr := chunk.SplitChEMBLTSVLimit(rc, in.RowsPerShard, in.MaxRows, func(index int, shard io.Reader) error {
 		art, err := domain.NewArtifact(job.ID, nil, domain.ArtifactShard,
 			fmt.Sprintf("shard-%d.tsv", index), in.ContentType, now)
 		if err != nil {
@@ -77,7 +86,7 @@ func (uc *SubmitDataset) Execute(ctx context.Context, in SubmitDatasetInput) (Su
 		putKeys = append(putKeys, art.StorageKey)
 		art.SetContent(ssum, ssize)
 
-		task, err := domain.NewShardTask(job.ID, index, in.Workload, art.ID, ssum, in.Parameters, 0, now)
+		task, err := domain.NewShardTask(job.ID, index, in.Workload, art.ID, ssum, in.Parameters, uc.maxAttempts, now)
 		if err != nil {
 			return err
 		}
@@ -88,7 +97,8 @@ func (uc *SubmitDataset) Execute(ctx context.Context, in SubmitDatasetInput) (Su
 	_ = rc.Close()
 	if splitErr != nil {
 		cleanup()
-		return SubmitDatasetResult{}, splitErr
+		// Dataset shape is caller input, not an internal coordinator failure.
+		return SubmitDatasetResult{}, domain.ErrInvalidInput
 	}
 
 	// 3. Persist job + all artifacts + all tasks atomically.
@@ -118,6 +128,76 @@ func (uc *SubmitDataset) Execute(ctx context.Context, in SubmitDatasetInput) (Su
 	}, nil
 }
 
+// validateUploadedWorkload is deliberately narrow until CTX-07/08/10 adds a
+// typed distributed-workload registry. In particular, running similarity-graph
+// independently per TSV shard is scientifically wrong: cross-shard pairs would
+// be absent from the apparent graph.
+func validateUploadedWorkload(workload string, parameters map[string]any) error {
+	if workload != "similarity-search" {
+		return domain.ErrInvalidInput
+	}
+	allowed := map[string]struct{}{
+		"query_smiles": {}, "top_k": {}, "threshold": {},
+		"threshold_direction": {}, "progress_every": {},
+	}
+	for key := range parameters {
+		if _, ok := allowed[key]; !ok {
+			return domain.ErrInvalidInput
+		}
+	}
+	query, ok := parameters["query_smiles"].(string)
+	if !ok || query == "" || len(query) > 200 {
+		return domain.ErrInvalidInput
+	}
+	if value, ok := parameters["top_k"]; ok && !isPositiveJSONInteger(value) {
+		return domain.ErrInvalidInput
+	}
+	if value, ok := parameters["progress_every"]; ok && !isNonNegativeJSONInteger(value) {
+		return domain.ErrInvalidInput
+	}
+	if value, ok := parameters["threshold"]; ok && !isUnitIntervalNumber(value) {
+		return domain.ErrInvalidInput
+	}
+	if value, ok := parameters["threshold_direction"]; ok && value != "greater" && value != "less" {
+		return domain.ErrInvalidInput
+	}
+	return nil
+}
+
+func isPositiveJSONInteger(value any) bool    { return isJSONInteger(value, false) }
+func isNonNegativeJSONInteger(value any) bool { return isJSONInteger(value, true) }
+
+func isJSONInteger(value any, allowZero bool) bool {
+	var n int64
+	switch v := value.(type) {
+	case int:
+		n = int64(v)
+	case int64:
+		n = v
+	case float64:
+		if math.Trunc(v) != v || v > math.MaxInt64 || v < math.MinInt64 {
+			return false
+		}
+		n = int64(v)
+	default:
+		return false
+	}
+	return n >= 0 && (allowZero || n > 0)
+}
+
+func isUnitIntervalNumber(value any) bool {
+	switch v := value.(type) {
+	case float64:
+		return !math.IsNaN(v) && !math.IsInf(v, 0) && v >= 0 && v <= 1
+	case int:
+		return v >= 0 && v <= 1
+	case int64:
+		return v >= 0 && v <= 1
+	default:
+		return false
+	}
+}
+
 // GetTaskInput resolves a task's input shard and opens it for streaming. The
 // caller closes the reader.
 type GetTaskInput struct {

coordinator/internal/usecase/usecase_test.go

@@ -11,6 +11,7 @@ import (
 
 	"github.com/google/uuid"
 
+	"github.com/emil28092005/SciMesh/coordinator/internal/authctx"
 	"github.com/emil28092005/SciMesh/coordinator/internal/domain"
 	"github.com/emil28092005/SciMesh/coordinator/internal/memstore"
 	"github.com/emil28092005/SciMesh/coordinator/internal/usecase"
@@ -40,6 +41,7 @@ type harness struct {
 	arts        *memstore.ArtifactRepo
 	blobs       *memstore.BlobStore
 	clk         *memstore.Clock
+	taskResults *memstore.TaskResultRepo
 
 	createJob   *usecase.CreateJob
 	submit      *usecase.SubmitDataset
@@ -55,6 +57,8 @@ type harness struct {
 	getInput    *usecase.GetTaskInput
 	expire      *usecase.ExpireLeases
 	cancel      *usecase.CancelJob
+	reduce      *usecase.ReduceJob
+	jobResult   *usecase.GetJobResult
 }
 
 func newHarness() *harness {
@@ -65,25 +69,105 @@ func newHarness() *harness {
 		arts:        memstore.NewArtifactRepo(),
 		blobs:       memstore.NewBlobStore(),
 		clk:         memstore.NewClock(time.Date(2026, 7, 21, 12, 0, 0, 0, time.UTC)),
+		taskResults: memstore.NewTaskResultRepo(),
 	}
 	tx := memstore.Tx{}
 	h.createJob = usecase.NewCreateJob(h.jobs, h.tasks, tx, h.clk)
-	h.submit = usecase.NewSubmitDataset(h.blobs, h.arts, h.jobs, h.tasks, tx, h.clk)
-	h.claim = usecase.NewClaimTask(h.tasks, h.clk, lease)
+	h.submit = usecase.NewSubmitDataset(h.blobs, h.arts, h.jobs, h.tasks, tx, h.clk, 3)
+	h.claim = usecase.NewClaimTask(h.tasks, h.jobs, h.work, tx, h.clk, lease)
 	h.renew = usecase.NewRenewLease(h.tasks, h.work, tx, h.clk, lease)
-	h.complete = usecase.NewCompleteTask(h.tasks, h.jobs, h.arts, tx, h.clk)
-	h.fail = usecase.NewFailTask(h.tasks, h.jobs, tx, h.clk)
+	h.complete = usecase.NewCompleteTask(h.tasks, h.jobs, h.arts, h.work, h.taskResults, tx, h.clk, 2)
+	h.fail = usecase.NewFailTask(h.tasks, h.jobs, h.work, tx, h.clk)
 	h.status = usecase.NewGetJobStatus(h.jobs, h.tasks)
 	h.results = usecase.NewListResults(h.tasks)
 	h.register = usecase.NewRegisterWorker(h.work, h.clk)
-	h.uploadArt = usecase.NewUploadArtifact(h.tasks, h.arts, h.blobs, h.clk)
+	h.uploadArt = usecase.NewUploadArtifact(h.tasks, h.work, h.arts, h.blobs, tx, h.clk)
 	h.downloadArt = usecase.NewDownloadArtifact(h.arts, h.blobs)
 	h.getInput = usecase.NewGetTaskInput(h.tasks, h.arts, h.blobs)
-	h.expire = usecase.NewExpireLeases(h.tasks, h.clk)
+	h.expire = usecase.NewExpireLeases(h.tasks, h.jobs, tx, h.clk)
 	h.cancel = usecase.NewCancelJob(h.jobs, h.tasks, tx, h.clk)
+	h.reduce = usecase.NewReduceJob(h.jobs, h.tasks, h.arts, h.blobs, tx, h.clk)
+	h.jobResult = usecase.NewGetJobResult(h.jobs, h.downloadArt)
 	return h
 }
 
+func TestSimilaritySearchReductionCreatesFinalArtifact(t *testing.T) {
+	h := newHarness()
+	jobID := h.seedJob(t, "similarity-search", 2)
+	_, err := h.jobs.Get(ctx, jobID)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if err := h.jobs.UpdateStatus(ctx, jobID, domain.JobRunning, nil); err != nil {
+		t.Fatal(err)
+	}
+	partials := []string{
+		"rank,chembl_id,canonical_smiles,similarity\n1,B,CCC,0.50000048\n",
+		"rank,chembl_id,canonical_smiles,similarity\n1,A,CC,0.50000049\n",
+	}
+	for _, partial := range partials {
+		taskID, attempt := h.leaseOne(t, "w1", "similarity-search")
+		art, err := h.uploadArt.Execute(ctx, usecase.UploadArtifactInput{TaskID: taskID, WorkerID: "w1", Attempt: attempt, Filename: "partial.csv", ContentType: "text/csv", Body: strings.NewReader(partial)})
+		if err != nil {
+			t.Fatal(err)
+		}
+		if _, err := h.complete.Execute(ctx, usecase.CompleteTaskInput{TaskID: taskID, WorkerID: "w1", Attempt: attempt, ResultArtifactID: art.ID}); err != nil {
+			t.Fatal(err)
+		}
+	}
+	if err := h.reduce.Execute(ctx, jobID); err != nil {
+		t.Fatal(err)
+	}
+	progress, err := h.status.Execute(ctx, jobID)
+	if err != nil || progress.Job.Status != domain.JobCompleted {
+		t.Fatalf("status=%s err=%v", progress.Job.Status, err)
+	}
+	art, body, err := h.jobResult.Execute(ctx, jobID)
+	if err != nil {
+		t.Fatal(err)
+	}
+	defer body.Close()
+	bytes, _ := io.ReadAll(body)
+	if art.Kind != domain.ArtifactFinalResult || string(bytes) != "rank,chembl_id,canonical_smiles,similarity\n1,A,CC,0.500000\n2,B,CCC,0.500000\n" {
+		t.Fatalf("unexpected final %q", bytes)
+	}
+}
+
+func TestSimilaritySearchReductionFailureIsSanitized(t *testing.T) {
+	h := newHarness()
+	jobID := h.seedJob(t, "similarity-search", 1)
+	if err := h.jobs.UpdateStatus(ctx, jobID, domain.JobRunning, nil); err != nil {
+		t.Fatal(err)
+	}
+	taskID, attempt := h.leaseOne(t, "w1", "similarity-search")
+	art, err := h.uploadArt.Execute(ctx, usecase.UploadArtifactInput{
+		TaskID: taskID, WorkerID: "w1", Attempt: attempt, Filename: "partial.csv",
+		ContentType: "text/csv", Body: strings.NewReader("rank,chembl_id,canonical_smiles,similarity\n2,A,CC,0.9\n"),
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if _, err := h.complete.Execute(ctx, usecase.CompleteTaskInput{
+		TaskID: taskID, WorkerID: "w1", Attempt: attempt, ResultArtifactID: art.ID,
+	}); err != nil {
+		t.Fatal(err)
+	}
+	if err := h.reduce.Execute(ctx, jobID); err != nil {
+		t.Fatal(err)
+	}
+	job, err := h.jobs.Get(ctx, jobID)
+	if err != nil {
+		t.Fatal(err)
+	}
+	if job.Status != domain.JobFailed || job.ErrorCode == nil || *job.ErrorCode != "reducer_failed" ||
+		job.ErrorMessage == nil || *job.ErrorMessage != "final result reduction failed" {
+		t.Fatalf("unexpected failed job: %+v", job)
+	}
+	if job.ResultArtifactID != nil {
+		t.Fatal("failed reduction must not expose a final result")
+	}
+}
+
 // seedJob creates a URI-chunked job with n chunks and returns its id.
 func (h *harness) seedJob(t *testing.T, workload string, n int) uuid.UUID {
 	t.Helper()
@@ -138,6 +222,44 @@ func TestClaimLeasesAndAdvancesAttempt(t *testing.T) {
 	}
 }
 
+func TestRegisteredWorkerCannotBroadenItsCapabilitiesAtClaim(t *testing.T) {
+	h := newHarness()
+	h.seedJob(t, "restricted", 1)
+	worker, err := h.register.Execute(ctx, usecase.RegisterWorkerInput{
+		Name: "search-only", Capabilities: []string{"similarity-search"},
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	claimed, err := h.claim.Execute(ctx, usecase.ClaimTaskInput{
+		WorkerID: worker.ID.String(), Workloads: []string{"restricted"},
+	})
+	if err != nil || claimed != nil {
+		t.Fatalf("claim = (%v, %v), want no compatible task", claimed, err)
+	}
+}
+
+func TestCreateJobRejectsUnsafeDistributedScientificPlans(t *testing.T) {
+	h := newHarness()
+	_, err := h.createJob.Execute(ctx, usecase.CreateJobInput{
+		Workload: "similarity-graph", InputURI: "s3://input",
+		Chunks: []usecase.ChunkInput{{ChunkIndex: 0, InputURI: "s3://chunk", InputSHA256: "sha"}},
+	})
+	if !errors.Is(err, domain.ErrInvalidInput) {
+		t.Errorf("graph job err = %v, want ErrInvalidInput", err)
+	}
+	_, err = h.createJob.Execute(ctx, usecase.CreateJobInput{
+		Workload: "similarity-search", InputURI: "s3://input", Parameters: map[string]any{"query_id": "CHEMBL1"},
+		Chunks: []usecase.ChunkInput{
+			{ChunkIndex: 0, InputURI: "s3://chunk0", InputSHA256: "sha"},
+			{ChunkIndex: 1, InputURI: "s3://chunk1", InputSHA256: "sha"},
+		},
+	})
+	if !errors.Is(err, domain.ErrInvalidInput) {
+		t.Errorf("sharded query_id job err = %v, want ErrInvalidInput", err)
+	}
+}
+
 func TestClaimEmptyQueueReturnsNil(t *testing.T) {
 	h := newHarness()
 	c, err := h.claim.Execute(ctx, usecase.ClaimTaskInput{WorkerID: "w1", Workloads: []string{"w"}})
@@ -146,6 +268,184 @@ func TestClaimEmptyQueueReturnsNil(t *testing.T) {
 	}
 }
 
+func TestRegisterWorkerDefaultsToTrusted(t *testing.T) {
+	h := newHarness()
+	// A shared-token registration carries no owner and no explicit trust.
+	w, err := h.register.Execute(ctx, usecase.RegisterWorkerInput{
+		Name: "lab", Capabilities: []string{"w"},
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if w.TrustLevel != domain.WorkerTrusted {
+		t.Errorf("trust = %q, want trusted", w.TrustLevel)
+	}
+	if w.OwnerID != nil {
+		t.Errorf("owner = %v, want nil for a shared-token worker", w.OwnerID)
+	}
+}
+
+func TestRegisterWorkerRecordsOwnerAndUntrusted(t *testing.T) {
+	h := newHarness()
+	owner := uuid.New()
+	w, err := h.register.Execute(ctx, usecase.RegisterWorkerInput{
+		Name: "volunteer", Capabilities: []string{"w"},
+		OwnerID: &owner, TrustLevel: domain.WorkerUntrusted,
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if w.TrustLevel != domain.WorkerUntrusted {
+		t.Errorf("trust = %q, want untrusted", w.TrustLevel)
+	}
+	if w.OwnerID == nil || *w.OwnerID != owner {
+		t.Errorf("owner = %v, want %v", w.OwnerID, owner)
+	}
+}
+
+func TestJWTCallerCannotClaimAsAnotherUsersWorker(t *testing.T) {
+	h := newHarness()
+	h.seedJob(t, "w", 1)
+
+	// A trusted lab worker owned by nobody (shared-token registration).
+	victim, _ := h.register.Execute(ctx, usecase.RegisterWorkerInput{Name: "lab", Capabilities: []string{"w"}})
+
+	// An attacker authenticated as a JWT user tries to claim as the lab worker.
+	attacker := authctx.With(ctx, authctx.Requester{UserID: uuid.New(), Role: "user"})
+	claimed, err := h.claim.Execute(attacker, usecase.ClaimTaskInput{WorkerID: victim.ID.String()})
+	if !errors.Is(err, domain.ErrWorkerNotFound) {
+		t.Fatalf("claim as another's worker = (%v, %v), want ErrWorkerNotFound", claimed, err)
+	}
+}
+
+func TestJWTCallerCannotMutateAnotherUsersWorkerLease(t *testing.T) {
+	h := newHarness()
+	h.seedJob(t, "w", 1)
+	victimOwner := uuid.New()
+	victim, err := h.register.Execute(ctx, usecase.RegisterWorkerInput{
+		Name: "victim", Capabilities: []string{"w"}, OwnerID: &victimOwner, TrustLevel: domain.WorkerTrusted,
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	claimed, err := h.claim.Execute(ctx, usecase.ClaimTaskInput{WorkerID: victim.ID.String()})
+	if err != nil || claimed == nil {
+		t.Fatalf("claim = (%v, %v)", claimed, err)
+	}
+	attacker := authctx.With(ctx, authctx.Requester{UserID: uuid.New(), Role: "user"})
+
+	if _, err := h.renew.Execute(attacker, usecase.RenewLeaseInput{TaskID: claimed.TaskID, WorkerID: victim.ID.String(), Attempt: claimed.Attempt}); !errors.Is(err, domain.ErrWorkerNotFound) {
+		t.Errorf("foreign heartbeat err = %v, want ErrWorkerNotFound", err)
+	}
+	if _, err := h.fail.Execute(attacker, usecase.FailTaskInput{TaskID: claimed.TaskID, WorkerID: victim.ID.String(), Attempt: claimed.Attempt, ErrorCode: "x"}); !errors.Is(err, domain.ErrWorkerNotFound) {
+		t.Errorf("foreign failure err = %v, want ErrWorkerNotFound", err)
+	}
+	if _, err := h.uploadArt.Execute(attacker, usecase.UploadArtifactInput{TaskID: claimed.TaskID, WorkerID: victim.ID.String(), Attempt: claimed.Attempt, Filename: "x.csv", ContentType: "text/csv", Body: strings.NewReader("x")}); !errors.Is(err, domain.ErrWorkerNotFound) {
+		t.Errorf("foreign upload err = %v, want ErrWorkerNotFound", err)
+	}
+	if _, err := h.complete.Execute(attacker, usecase.CompleteTaskInput{TaskID: claimed.TaskID, WorkerID: victim.ID.String(), Attempt: claimed.Attempt, ResultArtifactID: uuid.New()}); !errors.Is(err, domain.ErrWorkerNotFound) {
+		t.Errorf("foreign result err = %v, want ErrWorkerNotFound", err)
+	}
+}
+
+func TestJWTCallerClaimsAsOwnTrustedWorker(t *testing.T) {
+	h := newHarness()
+	h.seedJob(t, "w", 1)
+	owner := uuid.New()
+
+	// The user's own worker, trusted (e.g. a verified contributor).
+	mine, _ := h.register.Execute(ctx, usecase.RegisterWorkerInput{
+		Name: "mine", Capabilities: []string{"w"}, OwnerID: &owner, TrustLevel: domain.WorkerTrusted,
+	})
+
+	callerCtx := authctx.With(ctx, authctx.Requester{UserID: owner, Role: "user", Verified: true})
+	got, err := h.claim.Execute(callerCtx, usecase.ClaimTaskInput{WorkerID: mine.ID.String()})
+	if err != nil || got == nil {
+		t.Fatalf("own trusted worker claim = (%v, %v), want a task", got, err)
+	}
+}
+
+func TestUntrustedWorkerCanClaim(t *testing.T) {
+	h := newHarness()
+	h.seedJob(t, "w", 1)
+	owner := uuid.New()
+	worker, err := h.register.Execute(ctx, usecase.RegisterWorkerInput{
+		Name: "volunteer", Capabilities: []string{"w"},
+		OwnerID: &owner, TrustLevel: domain.WorkerUntrusted,
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	// Volunteers are no longer quarantined — they may claim; their results are
+	// gated by quorum at completion, not by withholding work.
+	claimed, err := h.claim.Execute(ctx, usecase.ClaimTaskInput{WorkerID: worker.ID.String()})
+	if err != nil || claimed == nil {
+		t.Fatalf("untrusted claim = (%v, %v), want a task", claimed, err)
+	}
+}
+
+// registerUntrusted registers a volunteer worker under a fresh owner.
+func (h *harness) registerUntrusted(t *testing.T, name, workload string) (*domain.Worker, uuid.UUID) {
+	t.Helper()
+	owner := uuid.New()
+	w, err := h.register.Execute(ctx, usecase.RegisterWorkerInput{
+		Name: name, Capabilities: []string{workload},
+		OwnerID: &owner, TrustLevel: domain.WorkerUntrusted,
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	return w, owner
+}
+
+func TestUntrustedResultNeedsQuorum(t *testing.T) {
+	h := newHarness()
+	jobID := h.seedJob(t, "w", 1)
+	if err := h.jobs.UpdateStatus(ctx, jobID, domain.JobRunning, nil); err != nil {
+		t.Fatal(err)
+	}
+	w1, _ := h.registerUntrusted(t, "v1", "w")
+	w2, _ := h.registerUntrusted(t, "v2", "w")
+
+	// First volunteer computes and submits — one vote, not yet quorum (2).
+	taskID, attempt := h.leaseOne(t, w1.ID.String(), "w")
+	art1 := h.uploadResult(t, taskID, w1.ID.String(), attempt)
+	if _, err := h.complete.Execute(ctx, usecase.CompleteTaskInput{TaskID: taskID, WorkerID: w1.ID.String(), Attempt: attempt, ResultArtifactID: art1}); err != nil {
+		t.Fatalf("first vote: %v", err)
+	}
+	if tk, _ := h.tasks.Get(ctx, taskID); tk.Status != domain.TaskPending {
+		t.Fatalf("after one vote status = %s, want pending", tk.Status)
+	}
+
+	// Second volunteer (distinct owner) computes the same bytes -> quorum -> done.
+	taskID2, attempt2 := h.leaseOne(t, w2.ID.String(), "w")
+	art2 := h.uploadResult(t, taskID2, w2.ID.String(), attempt2)
+	if _, err := h.complete.Execute(ctx, usecase.CompleteTaskInput{TaskID: taskID2, WorkerID: w2.ID.String(), Attempt: attempt2, ResultArtifactID: art2}); err != nil {
+		t.Fatalf("second vote: %v", err)
+	}
+	if tk, _ := h.tasks.Get(ctx, taskID); tk.Status != domain.TaskCompleted {
+		t.Fatalf("after quorum status = %s, want completed", tk.Status)
+	}
+}
+
+func TestTrustedResultCompletesDirectly(t *testing.T) {
+	h := newHarness()
+	jobID := h.seedJob(t, "w", 1)
+	if err := h.jobs.UpdateStatus(ctx, jobID, domain.JobRunning, nil); err != nil {
+		t.Fatal(err)
+	}
+	// A trusted (default) worker's single result completes the task immediately.
+	worker, _ := h.register.Execute(ctx, usecase.RegisterWorkerInput{Name: "lab", Capabilities: []string{"w"}})
+	taskID, attempt := h.leaseOne(t, worker.ID.String(), "w")
+	art := h.uploadResult(t, taskID, worker.ID.String(), attempt)
+	if _, err := h.complete.Execute(ctx, usecase.CompleteTaskInput{TaskID: taskID, WorkerID: worker.ID.String(), Attempt: attempt, ResultArtifactID: art}); err != nil {
+		t.Fatal(err)
+	}
+	if tk, _ := h.tasks.Get(ctx, taskID); tk.Status != domain.TaskCompleted {
+		t.Fatalf("trusted result status = %s, want completed", tk.Status)
+	}
+}
+
 func TestClaimRequiresWorkerID(t *testing.T) {
 	h := newHarness()
 	if _, err := h.claim.Execute(ctx, usecase.ClaimTaskInput{}); !errors.Is(err, domain.ErrInvalidInput) {
@@ -281,7 +581,7 @@ func TestUploadRejectsLeaseThatExpiresDuringStreaming(t *testing.T) {
 	h.seedJob(t, "w", 1)
 	taskID, attempt := h.leaseOne(t, "w1", "w")
 	h.uploadArt = usecase.NewUploadArtifact(
-		h.tasks, h.arts, expiringBlobStore{BlobStore: h.blobs, clock: h.clk}, h.clk,
+		h.tasks, h.work, h.arts, expiringBlobStore{BlobStore: h.blobs, clock: h.clk}, memstore.Tx{}, h.clk,
 	)
 
 	_, err := h.uploadArt.Execute(ctx, usecase.UploadArtifactInput{
@@ -414,6 +714,23 @@ func TestUploadArtifactRejectsForeignWorker(t *testing.T) {
 	}
 }
 
+func TestUploadArtifactIsIdempotentPerTaskAttempt(t *testing.T) {
+	h := newHarness()
+	h.seedJob(t, "w", 1)
+	taskID, attempt := h.leaseOne(t, "w1", "w")
+	first := h.uploadResult(t, taskID, "w1", attempt)
+	second, err := h.uploadArt.Execute(ctx, usecase.UploadArtifactInput{
+		TaskID: taskID, WorkerID: "w1", Attempt: attempt,
+		Filename: "retry.csv", ContentType: "text/csv", Body: strings.NewReader("different bytes"),
+	})
+	if err != nil {
+		t.Fatal(err)
+	}
+	if second.ID != first {
+		t.Errorf("retry artifact = %s, want existing %s", second.ID, first)
+	}
+}
+
 func TestDownloadArtifactRoundTrips(t *testing.T) {
 	h := newHarness()
 	h.seedJob(t, "w", 1)
@@ -434,10 +751,10 @@ func TestDownloadArtifactRoundTrips(t *testing.T) {
 
 func TestSubmitDatasetChunksAndServesInput(t *testing.T) {
 	h := newHarness()
-	tsv := "id\tsmiles\nA\tCC\nB\tCCC\nC\tCCCC\nD\tCCCCC\nE\tCCCCCC\n"
+	tsv := "chembl_id\tcanonical_smiles\nA\tCC\nB\tCCC\nC\tCCCC\nD\tCCCCC\nE\tCCCCCC\n"
 
 	res, err := h.submit.Execute(ctx, usecase.SubmitDatasetInput{
-		Workload: "w", RowsPerShard: 2, Filename: "chembl.tsv",
+		Workload: "similarity-search", Parameters: map[string]any{"query_smiles": "CCO"}, RowsPerShard: 2, Filename: "chembl.tsv",
 		ContentType: "text/tab-separated-values", Body: strings.NewReader(tsv),
 	})
 	if err != nil {
@@ -453,7 +770,7 @@ func TestSubmitDatasetChunksAndServesInput(t *testing.T) {
 		t.Errorf("job total = %d, want 3", prog.Total)
 	}
 
-	c, err := h.claim.Execute(ctx, usecase.ClaimTaskInput{WorkerID: "w1", Workloads: []string{"w"}})
+	c, err := h.claim.Execute(ctx, usecase.ClaimTaskInput{WorkerID: "w1", Workloads: []string{"similarity-search"}})
 	if err != nil || c == nil {
 		t.Fatalf("claim shard: %v", err)
 	}
@@ -472,9 +789,9 @@ func TestSubmitDatasetChunksAndServesInput(t *testing.T) {
 
 func TestSubmitDatasetLimitsRowsBeforeCreatingShards(t *testing.T) {
 	h := newHarness()
-	tsv := "id\tsmiles\nA\tCC\nB\tCCC\nC\tCCCC\nD\tCCCCC\nE\tCCCCCC\n"
+	tsv := "chembl_id\tcanonical_smiles\nA\tCC\nB\tCCC\nC\tCCCC\nD\tCCCCC\nE\tCCCCCC\n"
 	res, err := h.submit.Execute(ctx, usecase.SubmitDatasetInput{
-		Workload: "w", RowsPerShard: 2, MaxRows: 3, Filename: "chembl.tsv",
+		Workload: "similarity-search", Parameters: map[string]any{"query_smiles": "CCO"}, RowsPerShard: 2, MaxRows: 3, Filename: "chembl.tsv",
 		ContentType: "text/tab-separated-values", Body: strings.NewReader(tsv),
 	})
 	if err != nil {
@@ -485,6 +802,26 @@ func TestSubmitDatasetLimitsRowsBeforeCreatingShards(t *testing.T) {
 	}
 }
 
+func TestSubmitDatasetRejectsUnsupportedDistributedWorkloads(t *testing.T) {
+	h := newHarness()
+	_, err := h.submit.Execute(ctx, usecase.SubmitDatasetInput{
+		Workload: "similarity-graph", Parameters: map[string]any{"threshold": 0.7}, RowsPerShard: 2,
+		Filename: "chembl.tsv", ContentType: "text/tab-separated-values",
+		Body: strings.NewReader("chembl_id\tcanonical_smiles\nA\tCC\n"),
+	})
+	if !errors.Is(err, domain.ErrInvalidInput) {
+		t.Errorf("graph submission err = %v, want ErrInvalidInput", err)
+	}
+	_, err = h.submit.Execute(ctx, usecase.SubmitDatasetInput{
+		Workload: "similarity-search", Parameters: map[string]any{"query_id": "CHEMBL1"}, RowsPerShard: 2,
+		Filename: "chembl.tsv", ContentType: "text/tab-separated-values",
+		Body: strings.NewReader("chembl_id\tcanonical_smiles\nA\tCC\n"),
+	})
+	if !errors.Is(err, domain.ErrInvalidInput) {
+		t.Errorf("query_id submission err = %v, want ErrInvalidInput", err)
+	}
+}
+
 func TestCancelJobInvalidatesClaimedAndPendingTasks(t *testing.T) {
 	h := newHarness()
 	jobID := h.seedJob(t, "w", 3)
@@ -531,3 +868,22 @@ func TestExpireLeasesReclaims(t *testing.T) {
 		t.Errorf("expire = (%d, %v), want (1, nil)", n, err)
 	}
 }
+
+func TestFinalLeaseExpiryPersistsFailedJobAndCannotBeCancelled(t *testing.T) {
+	h := newHarness()
+	jobID := h.seedJob(t, "w", 1)
+	for attempt := 1; attempt <= domain.DefaultMaxAttempts; attempt++ {
+		h.leaseOne(t, "w1", "w")
+		h.clk.Advance(lease + time.Second)
+		if _, err := h.expire.Execute(ctx); err != nil {
+			t.Fatalf("expire attempt %d: %v", attempt, err)
+		}
+	}
+	progress, err := h.status.Execute(ctx, jobID)
+	if err != nil || progress.Job.Status != domain.JobFailed || progress.DeriveStatus() != domain.JobFailed {
+		t.Fatalf("progress = %+v, err = %v; want persisted failed job", progress, err)
+	}
+	if _, err := h.cancel.Execute(ctx, jobID); !errors.Is(err, domain.ErrJobNotCancellable) {
+		t.Errorf("cancel terminal lease failure = %v, want ErrJobNotCancellable", err)
+	}
+}

coordinator/internal/usecase/worker.go

@@ -22,6 +22,13 @@ func (uc *RegisterWorker) Execute(ctx context.Context, in RegisterWorkerInput) (
 	if err != nil {
 		return nil, err
 	}
+	w.OwnerID = in.OwnerID
+	// The transport layer resolves trust from the caller's credentials; fall
+	// back to the domain default (trusted) only when it was left unset, so a
+	// zero-value input never silently downgrades a shared-token worker.
+	if in.TrustLevel != "" {
+		w.TrustLevel = in.TrustLevel
+	}
 	if err := uc.workers.Insert(ctx, w); err != nil {
 		return nil, err
 	}

coordinator/internal/usecase/worker_authorization.go

@@ -0,0 +1,33 @@
+package usecase
+
+import (
+	"context"
+
+	"github.com/google/uuid"
+
+	"github.com/emil28092005/SciMesh/coordinator/internal/authctx"
+	"github.com/emil28092005/SciMesh/coordinator/internal/domain"
+)
+
+// authorizeWorkerOwner binds a JWT-authenticated requester to a worker. The
+// shared coordinator token intentionally has no requester and retains its
+// existing operator privileges.
+func authorizeWorkerOwner(ctx context.Context, workers WorkerRepository, workerID string) error {
+	requester, ok := authctx.From(ctx)
+	if !ok {
+		return nil
+	}
+	id, err := uuid.Parse(workerID)
+	if err != nil {
+		return domain.ErrWorkerNotFound
+	}
+	worker, err := workers.Get(ctx, id)
+	if err != nil {
+		return err
+	}
+	if worker.OwnerID == nil || *worker.OwnerID != requester.UserID {
+		// Mask ownership and existence from another user.
+		return domain.ErrWorkerNotFound
+	}
+	return nil
+}

coordinator/migrations/0009_unique_partial_result_attempt.down.sql

@@ -0,0 +1,5 @@
+BEGIN;
+
+DROP INDEX IF EXISTS uq_partial_result_task_attempt;
+
+COMMIT;

coordinator/migrations/0009_unique_partial_result_attempt.up.sql

@@ -0,0 +1,26 @@
+BEGIN;
+
+-- Old deployments can contain more than one partial result because earlier
+-- versions accepted repeated PUTs. Preserve the one referenced by a completed
+-- task and discard stale rows; unfinished tasks must upload again after a
+-- deploy, just as they do after a lost lease.
+DELETE FROM artifacts AS a
+USING tasks AS t
+WHERE a.task_id = t.id
+  AND a.kind = 'partial_result'::artifact_kind
+  AND t.status <> 'completed'::task_status;
+
+DELETE FROM artifacts AS a
+USING tasks AS t
+WHERE a.task_id = t.id
+  AND a.kind = 'partial_result'::artifact_kind
+  AND t.status = 'completed'::task_status
+  AND a.id <> t.result_artifact_id;
+
+-- One lease attempt has one durable partial result. This makes an upload retry
+-- idempotent and prevents repeated uploads from accumulating orphan artifacts.
+CREATE UNIQUE INDEX uq_partial_result_task_attempt
+    ON artifacts (task_id, attempt)
+    WHERE kind = 'partial_result'::artifact_kind;
+
+COMMIT;

coordinator/migrations/0010_job_reduction.down.sql

@@ -0,0 +1,7 @@
+BEGIN;
+
+ALTER TABLE jobs DROP COLUMN IF EXISTS error_message;
+ALTER TABLE jobs DROP COLUMN IF EXISTS error_code;
+ALTER TABLE jobs DROP COLUMN IF EXISTS reducer_started_at;
+
+COMMIT;

coordinator/migrations/0010_job_reduction.up.sql

@@ -0,0 +1,7 @@
+-- PostgreSQL enum values must be committed before they are used by a later
+-- transaction, so this migration intentionally has no BEGIN/COMMIT wrapper.
+ALTER TYPE job_status ADD VALUE IF NOT EXISTS 'reducing';
+
+ALTER TABLE jobs ADD COLUMN IF NOT EXISTS error_code text;
+ALTER TABLE jobs ADD COLUMN IF NOT EXISTS error_message text;
+ALTER TABLE jobs ADD COLUMN IF NOT EXISTS reducer_started_at timestamptz;

coordinator/migrations/0011_job_owner.down.sql

@@ -0,0 +1,6 @@
+BEGIN;
+
+DROP INDEX IF EXISTS ix_jobs_owner;
+ALTER TABLE jobs DROP COLUMN IF EXISTS owner_id;
+
+COMMIT;

coordinator/migrations/0011_job_owner.up.sql

@@ -0,0 +1,14 @@
+BEGIN;
+
+-- Who submitted this job. Equals users.id from the userservice, taken from the
+-- JWT `sub` claim. NOT a foreign key: users live in a separate service/database,
+-- so integrity is guaranteed by the signed token, not by the DB.
+--
+-- Nullable because rows created before auth existed have no owner; new inserts
+-- must supply it (enforced in the app, not the schema, during the MVP).
+ALTER TABLE jobs ADD COLUMN owner_id uuid;
+
+-- "List my jobs" / "admin filters by owner" scans by owner.
+CREATE INDEX ix_jobs_owner ON jobs (owner_id);
+
+COMMIT;

coordinator/migrations/0012_worker_trust.down.sql

@@ -0,0 +1,8 @@
+BEGIN;
+
+DROP INDEX IF EXISTS ix_workers_owner;
+ALTER TABLE workers DROP COLUMN IF EXISTS trust_level;
+ALTER TABLE workers DROP COLUMN IF EXISTS owner_id;
+DROP TYPE IF EXISTS worker_trust;
+
+COMMIT;

coordinator/migrations/0012_worker_trust.up.sql

@@ -0,0 +1,18 @@
+BEGIN;
+
+-- Whether a worker's results are accepted directly or must clear quorum.
+-- 'trusted'   — lab machine (shared token) or a verified/admin contributor.
+-- 'untrusted' — a plain enthusiast; results are quarantined until quorum (C2).
+CREATE TYPE worker_trust AS ENUM ('trusted', 'untrusted');
+
+-- Who registered this worker (userservice user id, from the JWT sub). NULL for
+-- workers registered with the shared service token. Not a foreign key: users
+-- live in a separate service/database.
+ALTER TABLE workers ADD COLUMN owner_id uuid;
+
+-- Existing rows were all shared-token lab workers, hence 'trusted'.
+ALTER TABLE workers ADD COLUMN trust_level worker_trust NOT NULL DEFAULT 'trusted';
+
+CREATE INDEX ix_workers_owner ON workers (owner_id);
+
+COMMIT;

coordinator/migrations/0013_task_results.down.sql

@@ -0,0 +1,5 @@
+BEGIN;
+
+DROP TABLE IF EXISTS task_results;
+
+COMMIT;

coordinator/migrations/0013_task_results.up.sql

@@ -0,0 +1,23 @@
+BEGIN;
+
+-- Quorum votes for a task computed by untrusted (volunteer) workers. A trusted
+-- worker's result completes the task directly and never lands here; an untrusted
+-- result is recorded as one vote, and the task is only completed once enough
+-- distinct owners submit the same result_sha256.
+--
+-- One vote per (task, owner): a single volunteer cannot stuff the ballot by
+-- running many workers under one account. A resubmission updates their vote.
+CREATE TABLE task_results (
+    task_id            uuid        NOT NULL REFERENCES tasks(id) ON DELETE CASCADE,
+    owner_id           uuid        NOT NULL,
+    result_sha256      text        NOT NULL,
+    result_artifact_id uuid        NOT NULL REFERENCES artifacts(id) ON DELETE CASCADE,
+    created_at         timestamptz NOT NULL DEFAULT now(),
+
+    PRIMARY KEY (task_id, owner_id)
+);
+
+-- Quorum check groups a task's votes by result_sha256.
+CREATE INDEX ix_task_results_quorum ON task_results (task_id, result_sha256);
+
+COMMIT;

coordinator/monitoring/grafana/dashboards/coordinator.json

@@ -0,0 +1,154 @@
+{
+  "annotations": { "list": [] },
+  "editable": true,
+  "graphTooltip": 1,
+  "schemaVersion": 39,
+  "tags": ["scimesh"],
+  "time": { "from": "now-15m", "to": "now" },
+  "refresh": "5s",
+  "title": "SciMesh Coordinator",
+  "uid": "scimesh-coordinator",
+  "panels": [
+    {
+      "type": "timeseries",
+      "title": "HTTP request rate by route",
+      "datasource": { "type": "prometheus", "uid": "prometheus" },
+      "gridPos": { "h": 8, "w": 12, "x": 0, "y": 0 },
+      "fieldConfig": { "defaults": { "unit": "reqps" }, "overrides": [] },
+      "targets": [
+        {
+          "refId": "A",
+          "datasource": { "type": "prometheus", "uid": "prometheus" },
+          "expr": "sum by (route) (rate(scimesh_http_requests_total[1m]))",
+          "legendFormat": "{{route}}"
+        }
+      ]
+    },
+    {
+      "type": "timeseries",
+      "title": "p95 latency by route",
+      "datasource": { "type": "prometheus", "uid": "prometheus" },
+      "gridPos": { "h": 8, "w": 12, "x": 12, "y": 0 },
+      "fieldConfig": { "defaults": { "unit": "s" }, "overrides": [] },
+      "targets": [
+        {
+          "refId": "A",
+          "datasource": { "type": "prometheus", "uid": "prometheus" },
+          "expr": "histogram_quantile(0.95, sum by (le, route) (rate(scimesh_http_request_duration_seconds_bucket[5m])))",
+          "legendFormat": "{{route}}"
+        }
+      ]
+    },
+    {
+      "type": "timeseries",
+      "title": "Requests by status",
+      "datasource": { "type": "prometheus", "uid": "prometheus" },
+      "gridPos": { "h": 8, "w": 12, "x": 0, "y": 8 },
+      "fieldConfig": { "defaults": { "unit": "reqps" }, "overrides": [] },
+      "targets": [
+        {
+          "refId": "A",
+          "datasource": { "type": "prometheus", "uid": "prometheus" },
+          "expr": "sum by (status) (rate(scimesh_http_requests_total[1m]))",
+          "legendFormat": "{{status}}"
+        }
+      ]
+    },
+    {
+      "type": "timeseries",
+      "title": "Goroutines",
+      "datasource": { "type": "prometheus", "uid": "prometheus" },
+      "gridPos": { "h": 8, "w": 6, "x": 12, "y": 8 },
+      "fieldConfig": { "defaults": { "unit": "short" }, "overrides": [] },
+      "targets": [
+        {
+          "refId": "A",
+          "datasource": { "type": "prometheus", "uid": "prometheus" },
+          "expr": "go_goroutines{job=\"coordinator\"}",
+          "legendFormat": "goroutines"
+        }
+      ]
+    },
+    {
+      "type": "timeseries",
+      "title": "Resident memory",
+      "datasource": { "type": "prometheus", "uid": "prometheus" },
+      "gridPos": { "h": 8, "w": 6, "x": 18, "y": 8 },
+      "fieldConfig": { "defaults": { "unit": "bytes" }, "overrides": [] },
+      "targets": [
+        {
+          "refId": "A",
+          "datasource": { "type": "prometheus", "uid": "prometheus" },
+          "expr": "process_resident_memory_bytes{job=\"coordinator\"}",
+          "legendFormat": "rss"
+        }
+      ]
+    },
+    {
+      "type": "row",
+      "title": "Domain state",
+      "collapsed": false,
+      "gridPos": { "h": 1, "w": 24, "x": 0, "y": 16 }
+    },
+    {
+      "type": "timeseries",
+      "title": "Tasks by status",
+      "datasource": { "type": "prometheus", "uid": "prometheus" },
+      "gridPos": { "h": 8, "w": 12, "x": 0, "y": 17 },
+      "fieldConfig": { "defaults": { "unit": "short", "custom": { "stacking": { "mode": "normal" }, "fillOpacity": 30 } }, "overrides": [] },
+      "targets": [
+        {
+          "refId": "A",
+          "datasource": { "type": "prometheus", "uid": "prometheus" },
+          "expr": "scimesh_tasks",
+          "legendFormat": "{{status}}"
+        }
+      ]
+    },
+    {
+      "type": "timeseries",
+      "title": "Jobs by status",
+      "datasource": { "type": "prometheus", "uid": "prometheus" },
+      "gridPos": { "h": 8, "w": 12, "x": 12, "y": 17 },
+      "fieldConfig": { "defaults": { "unit": "short", "custom": { "stacking": { "mode": "normal" }, "fillOpacity": 30 } }, "overrides": [] },
+      "targets": [
+        {
+          "refId": "A",
+          "datasource": { "type": "prometheus", "uid": "prometheus" },
+          "expr": "scimesh_jobs",
+          "legendFormat": "{{status}}"
+        }
+      ]
+    },
+    {
+      "type": "stat",
+      "title": "Queue depth (pending tasks)",
+      "datasource": { "type": "prometheus", "uid": "prometheus" },
+      "gridPos": { "h": 8, "w": 6, "x": 0, "y": 25 },
+      "fieldConfig": { "defaults": { "unit": "short", "color": { "mode": "thresholds" }, "thresholds": { "steps": [ { "color": "green", "value": null }, { "color": "yellow", "value": 50 }, { "color": "red", "value": 500 } ] } }, "overrides": [] },
+      "targets": [
+        {
+          "refId": "A",
+          "datasource": { "type": "prometheus", "uid": "prometheus" },
+          "expr": "sum(scimesh_tasks{status=\"pending\"})",
+          "legendFormat": "pending"
+        }
+      ]
+    },
+    {
+      "type": "timeseries",
+      "title": "Workers by status",
+      "datasource": { "type": "prometheus", "uid": "prometheus" },
+      "gridPos": { "h": 8, "w": 18, "x": 6, "y": 25 },
+      "fieldConfig": { "defaults": { "unit": "short", "custom": { "stacking": { "mode": "normal" }, "fillOpacity": 30 } }, "overrides": [] },
+      "targets": [
+        {
+          "refId": "A",
+          "datasource": { "type": "prometheus", "uid": "prometheus" },
+          "expr": "scimesh_workers",
+          "legendFormat": "{{status}}"
+        }
+      ]
+    }
+  ]
+}

coordinator/monitoring/grafana/provisioning/dashboards/provider.yml

@@ -0,0 +1,10 @@
+apiVersion: 1
+
+providers:
+  - name: SciMesh
+    type: file
+    disableDeletion: false
+    allowUiUpdates: true
+    options:
+      path: /var/lib/grafana/dashboards
+      foldersFromFilesStructure: false

coordinator/monitoring/grafana/provisioning/datasources/prometheus.yml

@@ -0,0 +1,10 @@
+apiVersion: 1
+
+datasources:
+  - name: Prometheus
+    uid: prometheus
+    type: prometheus
+    access: proxy
+    url: http://prometheus:9090
+    isDefault: true
+    editable: false

coordinator/monitoring/prometheus.yml

@@ -0,0 +1,10 @@
+# Prometheus scrape config for the SciMesh demo. Prometheus runs in the same
+# compose network as the coordinator, so it reaches it by service name.
+global:
+  scrape_interval: 5s
+  evaluation_interval: 5s
+
+scrape_configs:
+  - job_name: coordinator
+    static_configs:
+      - targets: ["coordinator:8080"]

coordinator/scripts/demo-ui.sh

@@ -0,0 +1,188 @@
+#!/usr/bin/env bash
+# Start a self-contained local UI demo with coordinator, PostgreSQL, and local
+# reference workers. It is intentionally for a developer's machine only.
+set -euo pipefail
+
+action=${1:-start}
+script_dir=$(CDPATH= cd -- "$(dirname -- "$0")" && pwd)
+coordinator_dir=$(CDPATH= cd -- "$script_dir/.." && pwd)
+repo_dir=$(CDPATH= cd -- "$coordinator_dir/.." && pwd)
+
+project=${DEMO_PROJECT:-scimesh-demo}
+postgres_port=${DEMO_POSTGRES_PORT:-55432}
+coordinator_port=${DEMO_COORDINATOR_PORT:-18080}
+userservice_port=${DEMO_USERSERVICE_PORT:-18081}
+prometheus_port=${DEMO_PROMETHEUS_PORT:-19090}
+grafana_port=${DEMO_GRAFANA_PORT:-13000}
+ui_token=${DEMO_UI_TOKEN:-demo-ui-secret}
+worker_token=${DEMO_WORKER_TOKEN:-demo-worker-token}
+# Shared HS256 secret; the coordinator verifies userservice tokens with it. Must
+# be at least 32 bytes (both services refuse a shorter one).
+jwt_secret=${DEMO_JWT_SECRET:-demo-jwt-secret-please-change-me-0123456789}
+# The first admin, seeded into the userservice on first boot.
+admin_email=${DEMO_ADMIN_EMAIL:-root@scimesh.local}
+admin_password=${DEMO_ADMIN_PASSWORD:-rootpassword}
+workers=${DEMO_WORKERS:-2}
+demo_dir=${DEMO_DIR:-.demo}
+case "$demo_dir" in
+  /*) ;;
+  *) demo_dir="$coordinator_dir/$demo_dir" ;;
+esac
+worker_bin=${SCIMESH_WORKER_BIN:-"$repo_dir/.venv/bin/scimesh-worker"}
+pid_file="$demo_dir/workers.pids"
+logs_dir="$demo_dir/logs"
+
+compose() {
+  POSTGRES_PORT="$postgres_port" \
+  COORDINATOR_PORT="$coordinator_port" \
+  USERSERVICE_PORT="$userservice_port" \
+  PROMETHEUS_PORT="$prometheus_port" \
+  GRAFANA_PORT="$grafana_port" \
+  UI_AUTH_TOKEN="$ui_token" \
+  WORKER_AUTH_TOKEN="$worker_token" \
+  JWT_SECRET="$jwt_secret" \
+  BOOTSTRAP_ADMIN_EMAIL="$admin_email" \
+  BOOTSTRAP_ADMIN_PASSWORD="$admin_password" \
+  docker compose -p "$project" \
+    -f "$coordinator_dir/docker-compose.yml" \
+    -f "$coordinator_dir/docker-compose.users.yml" \
+    -f "$coordinator_dir/docker-compose.monitoring.yml" "$@"
+}
+
+stop_workers() {
+  [[ -f "$pid_file" ]] || return 0
+  while IFS= read -r pid; do
+    [[ "$pid" =~ ^[0-9]+$ ]] || continue
+    command_line=$(ps -p "$pid" -o args= 2>/dev/null || true)
+    # Never kill a recycled PID or a worker launched outside this demo.
+    if [[ "$command_line" == *"$demo_dir/worker-"* ]]; then
+      kill "$pid" 2>/dev/null || true
+    fi
+  done < "$pid_file"
+  rm -f "$pid_file"
+}
+
+wait_for_coordinator() {
+  local attempt=0
+  until curl --fail --silent --show-error "http://localhost:$coordinator_port/health" >/dev/null; do
+    attempt=$((attempt + 1))
+    if (( attempt >= 45 )); then
+      echo "Coordinator did not become ready. Recent logs:" >&2
+      compose logs --tail=80 coordinator >&2 || true
+      exit 1
+    fi
+    sleep 1
+  done
+}
+
+wait_for_userservice() {
+  local attempt=0
+  until curl --fail --silent --show-error "http://localhost:$userservice_port/health" >/dev/null; do
+    attempt=$((attempt + 1))
+    if (( attempt >= 45 )); then
+      echo "Userservice did not become ready. Recent logs:" >&2
+      compose logs --tail=80 userservice >&2 || true
+      exit 1
+    fi
+    sleep 1
+  done
+}
+
+wait_for_workers() {
+  local attempt=0 registered overview cookie="$demo_dir/session.cookies"
+  # The dashboard API is behind a userservice session now, not basic auth. Log in
+  # as the seeded admin (who sees every worker) to obtain a session cookie.
+  curl --fail --silent -c "$cookie" \
+    --data-urlencode "email=$admin_email" \
+    --data-urlencode "password=$admin_password" \
+    "http://localhost:$coordinator_port/ui/login" >/dev/null 2>&1 || true
+  until false; do
+    overview=$(curl --fail --silent --show-error -b "$cookie" \
+      "http://localhost:$coordinator_port/ui/api/overview" 2>/dev/null || true)
+    # The overview contains no jobs at demo startup, so every `id` belongs to
+    # a registered worker. Avoid adding jq just for this local helper.
+    registered=$(printf '%s' "$overview" | grep -o '"id"' | wc -l | tr -d ' ' || true)
+    if [[ "$registered" =~ ^[0-9]+$ ]] && (( registered >= workers )); then
+      return 0
+    fi
+    attempt=$((attempt + 1))
+    if (( attempt >= 20 )); then
+      echo "Only $registered of $workers demo workers registered. Recent worker logs:" >&2
+      tail -n 40 "$logs_dir"/worker-*.log 2>/dev/null >&2 || true
+      exit 1
+    fi
+    sleep 1
+  done
+}
+
+start() {
+  if ! [[ "$workers" =~ ^[1-9][0-9]*$ ]]; then
+    echo "DEMO_WORKERS must be a positive integer (got $workers)." >&2
+    exit 2
+  fi
+  if [[ ! -x "$worker_bin" ]]; then
+    echo "Reference worker not found: $worker_bin" >&2
+    echo "Create it first from the repository root: python3 -m venv .venv && .venv/bin/pip install -e '.[dev]'" >&2
+    exit 2
+  fi
+  command -v docker >/dev/null || { echo "Docker is required." >&2; exit 2; }
+  command -v curl >/dev/null || { echo "curl is required." >&2; exit 2; }
+
+  stop_workers
+  mkdir -p "$logs_dir"
+  compose up -d --build
+  echo "Waiting for the coordinator on http://localhost:$coordinator_port ..."
+  wait_for_coordinator
+  echo "Waiting for the userservice on http://localhost:$userservice_port ..."
+  wait_for_userservice
+
+  : > "$pid_file"
+  for index in $(seq 1 "$workers"); do
+    work_dir="$demo_dir/worker-$index"
+    mkdir -p "$work_dir"
+    SCIMESH_COORDINATOR_URL="http://localhost:$coordinator_port" \
+    SCIMESH_BEARER_TOKEN="$worker_token" \
+    "$worker_bin" \
+      --worker-name "demo-worker-$index" \
+      --work-dir "$work_dir" \
+      >"$logs_dir/worker-$index.log" 2>&1 &
+    echo "$!" >> "$pid_file"
+  done
+  wait_for_workers
+
+  cat <<EOF
+
+SciMesh manual demo is ready.
+
+  UI:          http://localhost:$coordinator_port/ui   (shows a login page)
+  Admin login: $admin_email / $admin_password
+  Userservice: http://localhost:$userservice_port
+  Grafana:     http://localhost:$grafana_port   (anonymous view; admin/${GRAFANA_PASSWORD:-admin} to edit)
+  Prometheus:  http://localhost:$prometheus_port
+  Workers:     $workers local reference workers
+
+Sign in with the admin above, or register a new account from the login page.
+The admin sees every job; a plain user sees only their own. Upload a small
+ChEMBL TSV through “New similarity search”, then watch the job page update.
+Worker logs are in $logs_dir. Stop everything with:
+
+  make demo-down
+EOF
+}
+
+case "$action" in
+  start) start ;;
+  stop)
+    stop_workers
+    compose down
+    echo "SciMesh manual demo stopped."
+    ;;
+  logs)
+    echo "Worker logs: $logs_dir"
+    compose logs -f coordinator
+    ;;
+  *)
+    echo "Usage: $0 {start|stop|logs}" >&2
+    exit 2
+    ;;
+esac

coordinator/scripts/smoke.sh

@@ -51,6 +51,13 @@ check "register worker"                 201 -X POST "${HOST}/workers/register" "
 	-d '{"name":"smoke-worker","capabilities":["similarity_search"],"cpu_count":4,"memory_mb":8192}'
 check "register without capabilities → 400" 400 -X POST "${HOST}/workers/register" "${auth[@]}" \
 	-d '{"name":"bad"}'
+registration=$(curl -sS "${auth[@]}" -X POST "${HOST}/workers/register" \
+	-d '{"name":"smoke-worker-active","capabilities":["similarity_search","similarity-search"],"cpu_count":4}')
+worker_id=$(printf '%s' "$registration" | python3 -c 'import json,sys;print(json.load(sys.stdin)["worker_id"])' 2>/dev/null)
+if [[ -z "${worker_id:-}" ]]; then
+	echo "  ✗ could not register active worker: $registration"
+	exit 1
+fi
 
 echo
 echo "job lifecycle"
@@ -75,7 +82,7 @@ declare -A our_chunks
 task_id=""
 attempt=""
 for _ in $(seq 1 40); do
-	claim=$(curl -sS "${auth[@]}" -X POST "${HOST}/tasks/claim" -d '{"worker_id":"w1"}')
+	claim=$(curl -sS "${auth[@]}" -X POST "${HOST}/tasks/claim" -d "{\"worker_id\":\"${worker_id}\"}")
 	[[ -z "$claim" ]] && break # 204: queue drained
 
 	read -r c_job c_task c_chunk c_attempt < <(printf '%s' "$claim" |
@@ -100,7 +107,7 @@ else
 fi
 
 check "heartbeat"                          200 -X POST "${HOST}/tasks/${task_id}/heartbeat" "${auth[@]}" \
-	-d "{\"worker_id\":\"w1\",\"attempt\":${attempt}}"
+	-d "{\"worker_id\":\"${worker_id}\",\"attempt\":${attempt}}"
 
 # --- artifacts + result (uploads happen while the task is still leased) ---
 bearer=(-H "Authorization: Bearer ${TOKEN}")
@@ -108,43 +115,47 @@ bearer=(-H "Authorization: Bearer ${TOKEN}")
 # upload <filename> -> prints the artifact_id
 upload() {
 	curl -sS -X PUT "${HOST}/tasks/${task_id}/artifacts/$1" "${bearer[@]}" \
-		-H 'Content-Type: text/csv' -H 'X-Worker-ID: w1' -H "X-Task-Attempt: ${attempt}" \
+		-H 'Content-Type: text/csv' -H "X-Worker-ID: ${worker_id}" -H "X-Task-Attempt: ${attempt}" \
 		--data-binary $'query,match,score\nA,B,0.9\n' |
 		python3 -c 'import json,sys;print(json.load(sys.stdin)["artifact_id"])' 2>/dev/null
 }
 
 check "upload artifact"                    200 -X PUT "${HOST}/tasks/${task_id}/artifacts/result.csv" "${bearer[@]}" \
-	-H 'Content-Type: text/csv' -H 'X-Worker-ID: w1' -H "X-Task-Attempt: ${attempt}" \
+	-H 'Content-Type: text/csv' -H "X-Worker-ID: ${worker_id}" -H "X-Task-Attempt: ${attempt}" \
 	--data-binary $'query,match,score\nA,B,0.9\n'
 check "foreign worker upload → 409"        409 -X PUT "${HOST}/tasks/${task_id}/artifacts/x.csv" "${bearer[@]}" \
 	-H 'Content-Type: text/csv' -H 'X-Worker-ID: impostor' -H "X-Task-Attempt: ${attempt}" \
 	--data-binary 'x'
 
-# Two result artifacts, uploaded now while the lease is held: one to complete
-# with, a second to prove a different manifest is rejected after completion.
+# A retry of a PUT returns the same durable artifact for the task attempt.
 art_id=$(upload primary.csv)
 art_id2=$(upload secondary.csv)
+if [[ "$art_id" == "$art_id2" ]]; then
+	printf '  \033[32m✓\033[0m %-46s %s\n' "duplicate upload is idempotent" "$art_id"
+	pass=$((pass + 1))
+else
+	printf '  \033[31m✗\033[0m %-46s got %s and %s\n' "duplicate upload is idempotent" "$art_id" "$art_id2"
+	fail=$((fail + 1))
+fi
 check "download artifact"                  200 "${HOST}/artifacts/${art_id}/download" "${bearer[@]}"
 
 check "foreign worker submits → 409"       409 -X POST "${HOST}/tasks/${task_id}/result" "${auth[@]}" \
 	-d "{\"worker_id\":\"impostor\",\"attempt\":${attempt},\"result\":{\"artifact_id\":\"${art_id}\"}}"
 check "submit result"                      200 -X POST "${HOST}/tasks/${task_id}/result" "${auth[@]}" \
-	-d "{\"worker_id\":\"w1\",\"attempt\":${attempt},\"result\":{\"artifact_id\":\"${art_id}\"}}"
+	-d "{\"worker_id\":\"${worker_id}\",\"attempt\":${attempt},\"result\":{\"artifact_id\":\"${art_id}\"}}"
 check "replay same result → idempotent"    200 -X POST "${HOST}/tasks/${task_id}/result" "${auth[@]}" \
-	-d "{\"worker_id\":\"w1\",\"attempt\":${attempt},\"result\":{\"artifact_id\":\"${art_id}\"}}"
-check "different result → 409"             409 -X POST "${HOST}/tasks/${task_id}/result" "${auth[@]}" \
-	-d "{\"worker_id\":\"w1\",\"attempt\":${attempt},\"result\":{\"artifact_id\":\"${art_id2}\"}}"
+	-d "{\"worker_id\":\"${worker_id}\",\"attempt\":${attempt},\"result\":{\"artifact_id\":\"${art_id}\"}}"
 check "GET /jobs/{id}"                     200 "${HOST}/jobs/${job_id}" "${auth[@]}"
 
 echo
 echo "input validation"
 check "malformed uuid → 400"               400 -X POST "${HOST}/tasks/not-a-uuid/result" "${auth[@]}" \
-	-d '{"worker_id":"w1","attempt":1,"result":{"artifact_id":"00000000-0000-0000-0000-000000000000"}}'
+	-d "{\"worker_id\":\"${worker_id}\",\"attempt\":1,\"result\":{\"artifact_id\":\"00000000-0000-0000-0000-000000000000\"}}"
 # Note: Go's encoding/json matches field names case-insensitively, so
 # "worker_ID" would be accepted as "worker_id". Only a genuinely unknown key
 # trips DisallowUnknownFields.
 check "unknown json field → 400"           400 -X POST "${HOST}/tasks/claim" "${auth[@]}" \
-	-d '{"worker_id":"w1","totally_unknown":1}'
+	-d "{\"worker_id\":\"${worker_id}\",\"totally_unknown\":1}"
 check "unknown job → 404"                  404 "${HOST}/jobs/00000000-0000-0000-0000-000000000000" "${auth[@]}"
 
 echo
@@ -152,11 +163,11 @@ echo "dataset upload → chunking"
 # Upload a 5-row TSV split at 2 rows/shard → expect 3 shard tasks. The text
 # fields precede the file part, which the coordinator streams.
 up=$(curl -sS "${bearer[@]}" -X POST "${HOST}/jobs/upload" \
-	-F 'workload=similarity_search' \
-	-F 'parameters={"top_k":10}' \
+	-F 'workload=similarity-search' \
+	-F 'parameters={"query_smiles":"CCO","top_k":10}' \
 	-F 'chunk_rows=2' \
 	-F 'file=@-;filename=chembl.tsv;type=text/tab-separated-values' <<'TSV'
-id	smiles
+chembl_id	canonical_smiles
 A	CC
 B	CCC
 C	CCCC
@@ -180,7 +191,7 @@ fi
 up_input=""
 for _ in $(seq 1 30); do
 	c=$(curl -sS "${bearer[@]}" -H 'Content-Type: application/json' -X POST "${HOST}/tasks/claim" \
-		-d '{"worker_id":"up-w","capabilities":["similarity_search"]}')
+		-d "{\"worker_id\":\"${worker_id}\"}")
 	[[ -z "$c" ]] && break
 	cj=$(printf '%s' "$c" | python3 -c 'import json,sys;print(json.load(sys.stdin)["job_id"])' 2>/dev/null)
 	[[ "$cj" != "$up_job" ]] && continue

docs/api-contract.md

@@ -26,7 +26,8 @@ must be updated in the same change as any behaviour it describes.
 | `POST /tasks/{id}/heartbeat` | renew lease | ✅ done |
 | `POST /tasks/{id}/result` | complete | ✅ done, references `artifact_id` |
 | `POST /tasks/{id}/failure` | fail | ✅ done |
-| `GET /jobs/{id}` | progress | ✅ done |
+| `GET /jobs/{id}` | progress and final-result URI | ✅ done |
+| `GET /jobs/{id}/result` | download final CSV | ✅ done |
 | `PUT /tasks/{id}/artifacts/{name}` | upload partial | ✅ done |
 | `GET /artifacts/{id}/download` | download by id | ✅ done |
 | `POST /jobs/upload` | upload dataset, coordinator chunks it | ✅ done |
@@ -67,6 +68,35 @@ artifact. The coordinator splits the selected TSV rows into shard artifacts
 Each resulting task's claim response carries `input.uri = /tasks/{id}/input`,
 served by §5.4.
 
+## Job progress and final result
+
+```http
+GET /jobs/{job_id}
+Authorization: Bearer <token>
+```
+
+The response contains task counters and a derived status. A successful
+`similarity-search` enters `reducing` after the last shard completes, then
+becomes `completed` only after the coordinator stores its deterministic final
+CSV. At that point `result_uri` is present:
+
+```json
+{
+  "id": "uuid",
+  "status": "completed",
+  "total": 3,
+  "pending": 0,
+  "leased": 0,
+  "completed": 3,
+  "failed": 0,
+  "cancelled": 0,
+  "result_uri": "/jobs/uuid/result"
+}
+```
+
+`GET /jobs/{job_id}/result` downloads that final coordinator-owned CSV. Before
+the job is completed (or for a failed/cancelled job), it returns `404`.
+
 ## Stop a job
 
 ```http
@@ -107,7 +137,8 @@ Content-Type: application/json
 ```
 
 `cpu_count`/`memory_mb` are accepted for forward compatibility and not yet
-persisted. `capabilities` must be non-empty (an allowlisted workload set).
+persisted. `capabilities` must be non-empty. A claim uses the capabilities
+stored at registration; the request cannot broaden them.
 
 ## Claim task
 
@@ -120,7 +151,8 @@ Content-Type: application/json
 ```
 
 - `204 No Content`: no compatible task.
-- `200 OK`: a task is leased atomically.
+- `200 OK`: a task is leased atomically. `worker_id` must be a registered UUID;
+  its persisted capabilities, rather than this request field, decide eligibility.
 
 ```json
 {

docs/artifact-preview-task.md

@@ -0,0 +1,81 @@
+# Task: safely preview partial CSV artifacts in the UI
+
+## Assignment
+
+You are the junior developer implementing one contained UI feature: an
+authenticated operator can preview a small portion of a CSV artifact belonging
+to the job they are viewing. This is a diagnostic aid, not a final-results
+page.
+
+## Read first
+
+1. `AGENTS.md`
+2. `.agents/coordinator.md`
+3. `docs/web-interface-plan.md`
+4. `docs/api-contract.md`
+5. `coordinator/internal/transport/http/ui.go` and its tests
+
+## Current baseline
+
+The coordinator serves an authenticated local web UI. Job detail pages already
+list partial result artifacts and provide job-scoped downloads. The browser has
+no worker token and must not learn storage paths. A partial shard CSV is never
+a global or final molecular-search result.
+
+## Scope
+
+Add a **Preview** action next to eligible CSV artifacts on a job detail page.
+
+- Preview only artifacts that belong to the requested job.
+- Show at most the first **30 rows** and read at most **64 KiB** from storage.
+- State clearly when content was truncated.
+- Preserve the existing download action.
+- For a non-CSV artifact, return a friendly, sanitized explanation rather than
+  attempting to render bytes as text.
+- Use an existing UI route pattern or add a small UI-authenticated endpoint;
+  keep it separate from worker API routes.
+
+## Security rules
+
+- Require UI Basic Auth for every preview request.
+- Verify job ownership in the coordinator before opening the artifact; an
+  artifact ID from another job must not be previewable.
+- Never expose `storage_key`, filesystem paths, database errors, bearer tokens,
+  or worker-local information.
+- Do not use `innerHTML` for CSV fields. Use `html/template` escaping or
+  `textContent` so strings such as `<script>alert(1)</script>` are displayed as
+  data, not executed.
+- Do not load the complete artifact into memory.
+
+## Out of scope
+
+- Charts, molecule imagery, RDKit rendering, client-side CSV libraries, React,
+  and a new frontend service.
+- Changing job/task state, retrying tasks, or implementing reducer output.
+- Redesigning the broader dashboard or job-creation workflow; that belongs to
+  `docs/user-space-task.md`.
+
+## Suggested implementation shape
+
+Keep UI transport, use case, and storage responsibilities separate. Return a
+small view model containing artifact metadata, column headers, rows, and a
+`truncated` flag. Reuse existing coordinator-owned artifact access rather than
+reading a path supplied by the browser. Keep the handler streaming/limited.
+
+## Acceptance criteria
+
+- A valid partial CSV can be previewed from its own job detail page.
+- The result shows no more than 30 data rows and marks 64 KiB/row truncation.
+- Empty and malformed CSV content fail safely with a clear message.
+- A non-CSV artifact is rejected safely.
+- Unauthenticated access is rejected; a cross-job artifact request is not
+  disclosed or served.
+- HTML-like values are escaped in the rendered preview.
+- Existing artifact downloads still work.
+- Add Go tests for all cases above and run `go test ./...` and `go vet ./...`.
+
+## Handoff
+
+Work in one focused branch and one PR. Report files changed, any API impact,
+test commands/results, and known limitations. Do not commit datasets, generated
+CSV files, Docker volumes, `.venv`, or `worker-data/`.