Run the /btw API call as fire-and-forget in interactive mode so the
main conversation is not blocked while waiting for the answer. The
action now returns immediately after setting the pending item, and
the background promise updates the UI when the answer arrives.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Revert tools:[] to empty config {} to avoid provider compatibility
issues. Empty tools array is truthy and gets passed through to API
requests, which can cause errors on some providers. Omitting the
tools field entirely achieves the same effect (no tool access).
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Explicitly pass tools:[] to generateContent to prevent tool invocation
in side questions. Add model undefined guard for defensive safety.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Extract repeated error formatting into formatBtwError helper, remove
no-op marginTop={0}, and add comprehensive test coverage for all three
execution modes.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Remove redundant null check by passing the already-validated client and
model to the helper function instead of re-extracting them.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add ACP mode support with stream_messages async generator
- Add non-interactive mode support with simple message return
- Add wrap="wrap" to Text components for long text handling
- Extract askBtw helper to reduce duplication across execution modes
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Add pendingItem conflict guard to prevent overwriting other operations
- Use finally block to ensure setPendingItem(null) is always called
- Wrap "Thinking..." with t() for i18n support
- Remove unnecessary type assertion, use typed variable instead
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Allow users to ask quick "by the way" questions that use the current
conversation context but don't pollute the main conversation history.
Closes#2370
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Previously, `/memory show --project` and `/memory show --global` only checked
the first filename from `getCurrentGeminiMdFilename()` (i.e., `QWEN.md`),
ignoring other configured context files like `AGENTS.md`. This caused the
commands to report empty even when `AGENTS.md` existed and was being loaded
by the actual memory loading mechanism (`loadServerHierarchicalMemory`).
Changes:
- Replace `getCurrentGeminiMdFilename()` with `getAllGeminiMdFilenames()` in
memoryCommand.ts
- Add `findAllExistingMemoryFiles()` helper that iterates all configured
filenames and aggregates content from all existing files
- Update both `--project` and `--global` subcommands to display content from
all found memory files
- Add tests for fallback (only AGENTS.md exists) and dual-file scenarios
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
- Track user-initiated cancellation separately from failures
- Cancel round immediately when user denies a tool call
- Add message queue to handle input during streaming
- Add info messages during Arena operations (apply, stop, cleanup)
- Disable tmux/iTerm2 backends (only in-process mode supported)
- Polish UI: green tool count, updated warning prefix
This improves the Arena UX by providing clearer feedback and
properly handling user cancellations without treating them as failures.
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
Extract atomic file write utility into reusable module. Add arena system reminder injection so orchestrating agents can discover active arena sessions. Support in-process mode status file writing for external consumers.
This enables agent-to-agent collaboration where a parent agent can monitor and coordinate arena sessions via file-based status files.
The /export commands (html, md, json, jsonl) were incorrectly using
loadLastSession() which loads the last modified session from disk,
rather than the currently active session.
This fix uses config.getSessionId() to get the current session ID
and loads the correct session with loadSession(sessionId).
Fixes#2267
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
- Add AgentHeader component showing model, path, and git branch
- Separate modelId and modelName in RegisteredAgent for cleaner display
- Simplify worktree branch naming from worktrees/session/name to base-session-name
- Change loading text from "Agent is working…" to "Thinking…"
- Make agent footer always visible (not just when input is active)
This improves the agent collaboration UX by providing context about each
agent's environment and simplifies the git worktree management.
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
- Move useArenaInProcess from AppContainer to AgentViewProvider
- Replace polling with config.onArenaManagerChange() callback
- Add success-type progress messages when agents finish tasks
- Add isSuccessStatus helper for IDLE/COMPLETED status checks
- Reset input history position when arena session starts
This improves separation of concerns and eliminates the 500ms polling
interval in favor of immediate reactive updates when the arena manager
changes.
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
Add stripStartupContext to remove env-info from parent history and pass
chatHistory through ArenaManager → InProcessBackend → AgentInteractive →
AgentCore. This allows arena agents to start with conversational context
from the main session.
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
- Introduce AgentStatus.IDLE for agents that finished work but can accept follow-up messages
- Add isSettledStatus() helper to check if agent is settled (IDLE or terminal)
- Update ArenaManager to transition to IDLE after agents finish initial task
- Keep agent tabs visible when session is IDLE so users can continue interacting
- Fix listener cleanup to not detach on IDLE (agents remain alive)
- Update tests to expect 'idle' status after successful completion
This enables the arena collaboration feature where agents can receive
additional tasks after completing their initial work.
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
- Use 8-char short names derived from session UUID for worktrees
- Fix cleanup to use short worktreeDirName
- Simplify model display names (remove authType prefix)
- Improve messaging when <2 models available
- Show agent worktree paths in startup output
Prevents long path issues and provides clearer model setup guidance.
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
