Previous test used compressed 8k images (0.01MB) which was meaningless.
Now tests with actual large noise images that don't compress.
Realistic payload limits discovered:
- Anthropic: 6 x 3MB = ~18MB total (not 32MB as documented)
- OpenAI: 2 x 15MB = ~30MB total
- Gemini: 10 x 20MB = ~200MB total (very permissive)
- Mistral: 4 x 10MB = ~40MB total
- xAI: 1 x 20MB (strict request size limit)
- Groq: 5 x 5760px images (5 image + pixel limit)
- zAI: 2 x 15MB = ~30MB (50MB request limit)
- OpenRouter: 2 x 5MB = ~10MB total
Also fixed GEMINI_API_KEY env var (was GOOGLE_API_KEY).
Related to #120
Tested max 8kx8k images per provider:
- Anthropic: 100 (explicit limit, fails at 101)
- OpenAI: 100-200 (100 works, 200 times out)
- Mistral: 8 (explicit limit, fails at 9)
- xAI: 100-150 (100 works, 150 times out)
- Groq: 0 (8k exceeds 33M pixel limit)
- zAI: 400 (context window limited at 500)
- OpenRouter: 40 (context window limited at 50)
- Gemini: untested (no API key in test env)
Key finding: Anthropic's 'many images' rule does NOT cause API errors.
100 x 8kx8k images work fine. Anthropic likely auto-resizes internally.
Related to #120
Tests max image count, size, dimensions, and 8k stress test for:
- Anthropic, OpenAI, Gemini, Mistral, OpenRouter, xAI, Groq, zAI
Key finding: Anthropic's 'many images' rule (>20 images = 2000px max)
does NOT cause API errors. 100 x 8k images work fine. Anthropic likely
auto-resizes internally.
Related to #120
Tests provider-specific image limitations across all supported providers:
- Maximum number of images in context
- Maximum image size (bytes)
- Maximum image dimensions
Discovered limits (Dec 2025):
- Anthropic: 100 images, 5MB per image, 8000px max dimension
- OpenAI: 500 images, >=25MB per image
- Gemini: ~2500 images, >=40MB per image
- Mistral: 8 images, ~15MB per image
- OpenRouter: ~40 images (context limited), ~15MB per image
- Add AgentToolUpdateCallback type and optional onUpdate callback to AgentTool.execute()
- Add tool_execution_update event with toolCallId, toolName, args, partialResult
- Normalize tool_execution_end to always use AgentToolResult (no more string fallback)
- Bash tool streams truncated rolling buffer output during execution
- ToolExecutionComponent shows last N lines when collapsed (not first N)
- Interactive mode handles tool_execution_update events
- Update RPC docs and ai/agent READMEs
fixes#44
wrapTextWithAnsi() was processing each line independently after splitting
on newlines, losing ANSI state. When styled text contained embedded newlines
(e.g. from markdown paragraphs), subsequent lines would lose their styling.
Fixed by tracking ANSI state across lines and prepending active codes to
lines after the first.
Fixes#197
Previously, when reasoning was not specified, some providers like Gemini
with 'dynamic thinking' enabled by default would still use thinking.
Now explicitly sets thinkingEnabled: false (Anthropic) and
thinking: { enabled: false } (Google) when reasoning is undefined.
Closes#180
- Auto-enable all models after /login via POST /models/{model}/policy
- Use openai-responses API for gpt-5/o3/o4 models (not accessible via completions)
- Normalize tool call IDs when switching between github-copilot models with different APIs
(fixes#198: openai-responses generates 450+ char IDs with special chars that break other models)
- Update README with streamlined GitHub Copilot docs
- OAuth login for GitHub Copilot via /login command
- Support for github.com and GitHub Enterprise
- Models sourced from models.dev (Claude, GPT, Gemini, Grok, etc.)
- Dynamic base URL from token's proxy-ep field
- Use vscode-chat integration ID for API compatibility
- Documentation for model enablement at github.com/settings/copilot/features
Co-authored-by: cau1k <cau1k@users.noreply.github.com>
- Image component returns correct number of lines (rows) for TUI accounting
- Empty lines rendered first, then cursor moves up and image is drawn
- This clears the space the image occupies before rendering
- Add spacer before inline images in tool output
- Create ShowImagesSelectorComponent with borders like other selectors
- Use showSelector pattern for /show-images command
- Add terminal.showImages setting to settings-manager.ts
- Add /show-images slash command (only visible if terminal supports images)
- ToolExecutionComponent checks both terminal support and user setting
- Shows text fallback when inline images are disabled
- Add getCellDimensions/setCellDimensions to terminal-image.ts
- TUI queries cell size on startup via CSI 16 t
- Parse response and filter it from user input
- Invalidate and re-render when cell dimensions received
- Pass rows parameter to Kitty protocol for correct aspect ratio
coding-agent:
- Export loadSkillsFromDir, formatSkillsForPrompt, and related types
- Refactor skills.ts to expose public API
mom:
- Add skills auto-discovery from workspace/skills and channel/skills
- Fix skill loading to use host paths (not Docker container paths)
- Update README and system prompt with SKILL.md format docs
Breaking change: Pi skills must now be named SKILL.md inside a directory,
matching Codex CLI format. Previously any *.md file was treated as a skill.
Migrate by renaming ~/.pi/agent/skills/foo.md to ~/.pi/agent/skills/foo/SKILL.md
Track all thread message timestamps during a run. When response is
[SILENT], delete all thread messages before deleting the main message.
This prevents the 'This message was deleted' tombstone in Slack.
When mom responds with just [SILENT], the status message is deleted
and nothing is posted to Slack. Useful for periodic events that check
for activity but find nothing to report.
- Usage summary now shows context tokens vs model context window
- Updated CHANGELOG.md with all recent changes
- Updated README.md with new file structure (log.jsonl/context.jsonl)
- log.jsonl is source of truth, context.jsonl syncs from it at run start
- Backfill fetches missing messages from Slack API on startup
- Messages sent while mom is busy are logged and synced on next run
- Channel chatter (no @mention) logged but doesn't trigger processing
- Pre-startup messages (replayed by Slack) logged but not processed
- Stop command executes immediately, not queued
- Session header written immediately on new session creation
- Deduplicate messages by timestamp
- Strip @mentions from backfilled messages
- Remove old slack.ts and main.ts, rename *-new.ts versions
- Remove user message logging from agent.ts (Slack handler already logs it)
- Add excludeTimestamp param to syncFromLog to skip current @mention
- Sync happens once before each prompt with current message excluded
- Ensures channel chatter and backfilled messages are in context
- Syncs on session creation and before each subsequent run
- User messages formatted as '[username]: text' in context
- Export AgentSession, SessionManager, SettingsManager, compaction from coding-agent
- Create MomSessionManager for channel-based context.jsonl storage
- Create MomSettingsManager for mom-specific settings
- Refactor agent.ts to use AgentSession instead of ephemeral Agent
- Split logging: tool results go to context.jsonl, human messages to log.jsonl
- Enable auto-compaction and overflow detection from coding-agent
Part of #115
Mario Zechner