When daemon sources kugetsu-session.sh to call cmd_start/cmd_continue,
it needs access to functions from kugetsu-config.sh, kugetsu-index.sh,
kugetsu-worktree.sh, and kugetsu-log.sh. Add sourcing at top of
kugetsu-session.sh.
The daemon (kugetsu-queue-daemon.sh) sources kugetsu-index.sh but not the main kugetsu script.
Move update_queue_item_state and kugetsu_add_notification to kugetsu-index.sh
so the daemon can use these functions when processing tasks.
When cmd_delegate detects an issue ref with number (e.g. git.fbrns.co/shoko/kugetsu#158),
it was calling cmd_start directly which tries to create worktree and clone.
This breaks the queue-based workflow where daemon should handle task execution.
Now cmd_delegate calls enqueue_task to add to queue, and daemon processes
tasks by calling cmd_start/cmd_continue as appropriate.
- Move issue_ref_to_filename and filename_to_issue_ref to kugetsu-index.sh
(where they logically belong, instead of in main kugetsu script)
- Refactor queue daemon to use cmd_start/cmd_continue for session management
- Daemon now checks if worktree/session exists → cmd_continue, else → cmd_start
- Removes ~40 lines of direct opencode session forking logic from daemon
- cmd_start/cmd_continue handle worktree creation, session forking, and tracking
This simplifies the daemon significantly and centralizes session management
in kugetsu-session.sh where it belongs.
- Load GITEA_TOKEN from ~/.kugetsu/env/default.env at daemon startup
- Use --fork --session --dir instead of --continue to run in correct directory
- Create worktree if it doesn't exist for the issue
- Track forked session ID (not parent pm_session) for completion detection
- Forked session ends when task completes, parent pm_session continues
Fixes#156
1. cmd_status now validates session IDs against opencode session list
- Reports 'error: base session X not found in opencode' if missing
- Reports 'error: pm_agent session X not found in opencode' if missing
2. Test suite now uses isolated KUGETSU_DIR=/tmp/test-kugetsu-$$
- All tests use separate test directory instead of ~/.kugetsu
- Prevents test suite from corrupting real user data
- Cleanup removes test directory entirely
Fixes#148
Add third pattern to parse_issue_ref_from_message() to support the mixed
format 'gitserver/owner/repo#number' (e.g., git.fbrns.co/shoko/kugetsu#116).
Previously only two formats were supported:
1. Full URL: #116
2. Short format: shoko/kugetsu#116
Now supports:
3. Mixed format: git.fbrns.co/shoko/kugetsu#116
Fixes#144
Split the monolithic kugetsu script into modular components:
Modules created:
- kugetsu-config.sh - Config/env loading and global variables
- kugetsu-index.sh - Index.json read/write via JSON
- kugetsu-worktree.sh - Git worktree operations
- kugetsu-log.sh - Structured logging and notifications
- kugetsu-session.sh - Session create/fork/destroy logic
- kugetsu-queue-daemon.sh - Queue daemon subprocess
Main script (kugetsu) is now a thin dispatcher that sources all modules.
Acceptance Criteria:
- All existing commands work exactly as before
- Main script sources modules
- Each module is independently testable
Fixes#116
Fixes#142 - process_queue silently skips all queue items because
issue_ref and message Python extraction commands were missing a closing
parenthesis. The error was silently swallowed by 2>/dev/null causing
both variables to be empty, so every queue item was skipped.
Implements #137 - Agent timeout handling.
Changes:
- Add TASK_TIMEOUT_HOURS config (default: 1 hour)
- Update queue item to track opencode_session_id and pid
- Add check_task_timeouts() function that:
- Checks notified tasks against timeout threshold
- Kills process if exceeded
- Marks session as 'timeout' state
- Integrate timeout check into queue daemon loop
Timeout behavior:
- Task is marked 'notified' when PM receives it
- If not completed within TASK_TIMEOUT_HOURS, task is killed
- Queue item marked 'error', session marked 'timeout'
- Extract hostname from pr_url instead of hardcoding domains
- Look up server base URL from GIT_SERVERS config
- Append /api/v1 to derive API URL (configurable per server)
- Works with any server configured in GIT_SERVERS
- Rename update-pr to set-pr for clarity (it's setting the PR URL, not updating PR)
- Add optional pr-url argument to kugetsu start command
Usage: kugetsu start <issue-ref> <message> [pr-url]
- If pr-url is provided at start, it's stored directly in session file
- Add WORKTREE_CHECK_PR_STATUS config (default: true)
- Add pr_url and branch_name fields to session files
- Add check_pr_status() to query PR status via API (Gitea/GitHub)
- Add update_session_pr_url() to update PR URL in session
- Add kugetsu update-pr command to set PR URL
- Modify cmd_destroy to check PR status before destroying worktree
Closes#135
- Parse issue refs from message (gitserver.com/owner/repo/issues/123 or owner/repo#123)
- Find existing worktrees/sessions by issue number
- Ask user to confirm which worktree to use, or delegate anyway
- Inject missing info context to PM agent
- Inject selected worktree context to PM agent
Fixes#128
The --dir flag only sets directory for the subprocess, not the session's
stored directory in opencode's SQLite DB. This was already fixed for
cmd_start in v0.1.10, but cmd_continue still had the bug.
Fixes#127
1. Init: cd to ~/.kugetsu-worktrees before creating base session
This keeps all worktrees inside a predictable directory structure
and avoids external_directory permission issues
2. Init: Clear old logs but keep repos.json, config, and env files
3. Fork context: Add kugetsu_get_fork_context() that provides:
- Important working rules (stop on error, don't pivot)
- Repository configuration from repos.json
- Environment file location info
4. Fork message: Prepend context to user message when forking session
Previously destroy only removed local session files but didn't delete
the sessions from opencode's database. This caused init to reuse the
same session with old context.
Now destroy calls 'opencode session delete <id>' to properly remove
the session from opencode.
Warn users if running kugetsu init from a directory with files or
git repository. This prevents project context from contaminating the
base session, which causes forked sessions to have unwanted context.
opencode session list doesn't show sessions in ~/.kugetsu-worktrees/ directories.
This caused detection to fail even though sessions were being created.
Now we query the database directly for sessions matching the worktree path.
Also fixed database path in fix_session_permissions (was ~/.opencode/, should be ~/.local/share/opencode/).
Previously we only fixed base session permissions before forking.
But permissions are NOT inherited from parent to child.
Now we update the newly created session's permissions immediately
after detection, ensuring the forked session can access external
directories like ~/.kugetsu/worktrees/.
- Call fix_session_permissions in cmd_start before forking to ensure
base session has correct permissions for external_directory access
- Add debug logging to show forked session's directory and permissions
after creation to help diagnose permission inheritance issues
1. Move session detection BEFORE checking if fork process is still running.
Previous code broke out of loop if forked process exited, skipping detection.
2. Add database query debugging when detection fails to help diagnose
why opencode session list might miss newly created sessions.
- Capture fork output to log file for debugging
- Track fork PID to detect if process exits early
- Retry session detection up to 10 seconds instead of 1 second
- Show fork log output when session creation fails
- Improve error message to indicate timeout
Issue #105: opencode run --fork/--continue --dir <path> fails to create sessions
Root cause: The --dir flag breaks session creation in opencode. Sessions
fail to be created when --dir is used with --fork or --continue.
Solution: Instead of using --dir flag, create worktrees inside the parent
session's directory and use 'cd $worktree_path && opencode run ...' to
change directory before running opencode.
Key changes:
- Worktrees now created at $PWD/.kugetsu-worktrees/{issue-ref}/ instead
of $WORKTREES_DIR/{issue-ref}/
- .kugetsu-worktrees is a hidden directory (git ignored by default)
- cmd_start and cmd_continue now use 'cd && opencode run' instead of
'opencode run --dir'
This approach works because:
1. Worktree is inside parent's directory tree (permission granted)
2. cd properly changes working directory before opencode runs
3. Session gets created with correct directory set
4. No .gitignore entry needed (. prefix makes it hidden from git)
- Test E7: verify fix_session_permissions function exists
- Test E8: verify cmd_doctor --fix-permissions flag is recognized
- Test E9: verify permission JSON is valid JSON
- Test E10: verify SQL UPDATE syntax works correctly
These tests verify the fix without requiring actual opencode installation.
- Add test E7: verify fix_session_permissions function exists
- Add test E8: verify cmd_doctor --fix-permissions flag is recognized
- Add fix_session_permissions call to cmd_init to set permissions
when initializing new sessions
Add --fix-permissions flag to cmd_doctor:
kugetsu doctor --fix-permissions
The fix_session_permissions() function:
- Updates base session and PM agent session permissions in SQLite
- Sets external_directory pattern to '*' with action 'allow'
- This fixes the issue where PM agent cannot access external directories
This addresses issue #36 where PM agent external_directory permission fails.
Fixes#36
Export KUGETSU_TEMP_DIR in cmd_delegate so subagents can use it
instead of /tmp which may be blocked by opencode.
Default: ~/.local/share/opencode/tool-output
This allows agents to write temp files in an allowed directory
instead of /tmp which is blocked in headless mode.
Fixes#73
Add tests for env pass-through feature:
- Test env command exists and lists files
- Test env set creates env file
- Test env show masks sensitive values (GITEA_TOKEN)
- Test set -a exports variables to child processes
- Test pm-agent.env takes precedence over default.env
- Test cmd_init creates env template files
These tests ensure the env pass-through mechanism works correctly
and that variables are properly exported to subagents.
