The 100ms timeout was causing Kitty protocol detection to fail when the
terminal response was delayed (e.g., due to event loop blocking during
startup). This resulted in shift+enter not working in some scenarios.
Changes:
- Remove timeout-based Kitty detection, process input immediately
- Detect Kitty response in stdinBuffer output (handles split data)
- Add modifyOtherKeys fallback for terminals without Kitty support
(matches xterm format \x1b[27;modifier;keycode~)
- New /models command with toggle UI for each available model
- Changes persist to enabledModels in settings.json
- Updates take effect immediately for Ctrl+P cycling
- Added documentation to README.md
closes#626
Fires when the model changes via /model command, model cycling (Ctrl+P),
or session restore. Includes source field and previousModel.
Add model-status.ts example extension demonstrating status bar updates.
closes#628
- Skills registered as /skill:name commands for quick access
- Toggle via /settings or skills.enableSkillCommands in settings.json
- Fuzzy matching for all slash command autocomplete (type /skbra for /skill:brave-search)
- Moved fuzzy module from coding-agent to tui package for reuse
Closes#630 by @Dwsy (reimplemented with fixes)
- Extracts user/assistant messages from session files for a given cwd
- Splits into ~100k char files to fit in context
- Spawns pi subagents to analyze each file for recurring patterns
- Compares against existing AGENTS.md to mark NEW vs EXISTING patterns
- Final aggregation step creates FINAL-SUMMARY.txt with consolidated findings
- Shows progress with tool call args during analysis
Removes MomSessionManager which was missing methods expected by AgentSession
(appendMessage, getBranch, etc.), causing runtime crashes.
Now uses the standard SessionManager from coding-agent with a fixed context.jsonl
path per channel. The syncLogToSessionManager function handles syncing messages
from log.jsonl using SessionManager's public API.
Ref #595
- Send SIGWINCH to self on terminal start to refresh stale dimensions (Unix only)
- Change requestRender(true) to set previousWidth = -1 to trigger widthChanged
- Update first render condition to skip when widthChanged is true
Fixes#599
Applied streaming TextDecoder fix to executeBashWithOperations(), matching
the fix in executeBash() from PR #433. Remote execution (SSH, containers)
was still using Buffer.toString() which corrupts multi-byte UTF-8 sequences
split across chunk boundaries.
fixes#608
Renamed from supportsStreamOptions to clarify this controls stream_options: { include_usage: true }.
Defaults to true (no behavioral change for existing providers).
Providers like gatewayz.ai that reject this parameter can set supportsUsageInStreaming: false in model config.
Based on #596 by @XesGaDeus
OpenRouter with models like openai/gpt-5.2 enforces strict mode which
requires all properties in the required array. Setting strict: false
allows optional parameters without null unions, matching the approach
already used in openai-responses.ts.
- /model <search> pre-filters selector or auto-selects on exact match
- Support provider/model syntax for disambiguation (e.g., /model openai/gpt-4)
- Auto-select logic moved to InteractiveMode to avoid selector UI flicker
Closes#587
Demonstrates running interactive commands (vim, git rebase, htop, etc.)
with full terminal access using user_bash event and ctx.ui.custom().
- Auto-detects interactive commands from built-in list
- !i prefix to force interactive mode
- Configurable via INTERACTIVE_COMMANDS/INTERACTIVE_EXCLUDE env vars
closes#532
The --no-skills flag set options.skills = [] in main.ts, but the
interactive mode UI would rediscover skills anyway because it called
loadSkills() directly instead of using the already-loaded skills.
Changes:
- Add AgentSession.skills and AgentSession.skillWarnings properties
- discoverSkills() now returns { skills, warnings } instead of Skill[]
- Interactive mode uses session.skills instead of calling loadSkills()
- Update SDK docs and examples for new return type
Fixes#577
Adds overlay rendering capability to the TUI, enabling floating modal
components that render on top of existing content without clearing the screen.
- Add showOverlay(), hideOverlay(), hasOverlay() methods to TUI
- Implement ANSI-aware line compositing via extractSegments()
- Support overlay stack (multiple overlays, later on top)
- Add { overlay: true } option to ctx.ui.custom()
- Add overlay-test.ts example extension
Also fixes pre-existing bug where bash tool output cached visual lines
at fixed terminal width, causing crashes on terminal resize.
Co-authored-by: Nico Bailon <nico.bailon@gmail.com>
- user_bash event for intercepting ! and !! commands (#528)
- Extensions can return { operations } or { result } to redirect/replace
- executeBashWithOperations() for custom BashOperations execution
- session.recordBashResult() for extensions handling bash themselves
- Theme API: getAllThemes(), getTheme(), setTheme() on ctx.ui
- mac-system-theme.ts example: sync with macOS dark/light mode
- Updated ssh.ts to use user_bash event
- Remove <thinking> tag generation from google-shared.ts, transorm-messages.ts, openai-completions.ts
- Thinking blocks now convert to plain text when switching models (prevents models mimicking tags)
- Skip empty thinking blocks to avoid API errors
- Keep thinking blocks only when same provider AND same model
fixes#561
When passing a string systemPrompt to createAgentSession(), it is now
used as-is without appending context files and skills. This matches the
documented behavior: 'String replaces default, function receives default
and returns final.'
Previously, string systemPrompt would have context files and skills
appended, causing duplication if they were already in the string.
fixes#543
Mario Zechner