cancel

SKILL.md

Cancel Skill

Intelligent cancellation that detects and cancels the active OMC mode.

The cancel skill is the standard way to complete and exit any OMC mode.
When the stop hook detects work is complete, it instructs the LLM to invoke
this skill for proper state cleanup. If cancel fails or is interrupted,
retry with --force flag, or wait for the 2-hour staleness timeout as
a last resort.

What It Does
Automatically detects which mode is active and cancels it:

Autopilot: Stops workflow, preserves progress for resume

Ralph: Stops persistence loop, clears linked ultrawork if applicable

Ultrawork: Stops parallel execution (standalone or linked)

UltraQA: Stops QA cycling workflow

Swarm: Stops coordinated agent swarm, releases claimed tasks

Ultrapilot: Stops parallel autopilot workers

Pipeline: Stops sequential agent pipeline

Team: Sends shutdown_request to all teammates, waits for responses, calls TeamDelete, clears linked ralph if present

Team+Ralph (linked): Cancels team first (graceful shutdown), then clears ralph state. Cancelling ralph when linked also cancels team first.

Usage

/oh-my-claudecode:cancel

Or say: "cancelomc", "stopomc"

Critical: Deferred Tool Handling

The state management tools (state_clear, state_read, state_write, state_list_active,
state_get_status) may be registered as deferred tools by Claude Code. Before calling
any state tool, you MUST first load all of them via ToolSearch:

ToolSearch(query="select:mcp__plugin_oh-my-claudecode_t__state_clear,mcp__plugin_oh-my-claudecode_t__state_read,mcp__plugin_oh-my-claudecode_t__state_write,mcp__plugin_oh-my-claudecode_t__state_list_active,mcp__plugin_oh-my-claudecode_t__state_get_status")

If state_clear is unavailable or fails, use this bash fallback as an emergency
escape from the stop hook loop. This is NOT a full replacement for the cancel flow —
it only removes state files to unblock the session. Linked modes (e.g. ralph→ultrawork,
autopilot→ralph/ultraqa) must be cleared separately by running the fallback once per mode.

Replace MODE with the specific mode (e.g. ralplan, ralph, ultrawork, ultraqa).

WARNING: Do NOT use this fallback for autopilot or omc-teams. Autopilot requires
state_write(active=false) to preserve resume data. omc-teams requires tmux session
cleanup that cannot be done via file deletion alone.

# Fallback: direct file removal when state_clear MCP tool is unavailable
SESSION_ID="${CLAUDE_SESSION_ID:-${CLAUDECODE_SESSION_ID:-}}"
REPO_ROOT="$(git rev-parse --show-toplevel 2>/dev/null || { d="$PWD"; while [ "$d" != "/" ] && [ ! -d "$d/.omc" ]; do d="$(dirname "$d")"; done; echo "$d"; })"

# Cross-platform SHA-256 (macOS: shasum, Linux: sha256sum)
sha256portable() { printf '%s' "$1" | (sha256sum 2>/dev/null || shasum -a 256) | cut -c1-16; }

# Resolve state directory (supports OMC_STATE_DIR centralized storage)
if [ -n "${OMC_STATE_DIR:-}" ]; then
  # Mirror getProjectIdentifier() from worktree-paths.ts
  SOURCE="$(git remote get-url origin 2>/dev/null || echo "$REPO_ROOT")"
  HASH="$(sha256portable "$SOURCE")"
  DIR_NAME="$(basename "$REPO_ROOT" | sed 's/[^a-zA-Z0-9_-]/_/g')"
  OMC_STATE="$OMC_STATE_DIR/${DIR_NAME}-${HASH}/state"
  [ ! -d "$OMC_STATE" ] && { echo "ERROR: State dir not found at $OMC_STATE" >&2; exit 1; }
elif [ "$REPO_ROOT" != "/" ] && [ -d "$REPO_ROOT/.omc" ]; then
  OMC_STATE="$REPO_ROOT/.omc/state"
else
  echo "ERROR: Could not locate .omc state directory" >&2
  exit 1
fi
MODE="ralplan"  # <-- replace with the target mode

# Clear session-scoped state for the specific mode
if [ -n "$SESSION_ID" ] && [ -d "$OMC_STATE/sessions/$SESSION_ID" ]; then
  rm -f "$OMC_STATE/sessions/$SESSION_ID/${MODE}-state.json"
  rm -f "$OMC_STATE/sessions/$SESSION_ID/${MODE}-stop-breaker.json"
  rm -f "$OMC_STATE/sessions/$SESSION_ID/skill-active-state.json"
  # Write cancel signal so stop hook detects cancellation in progress
  NOW_ISO="$(date -u +"%Y-%m-%dT%H:%M:%SZ")"
  EXPIRES_ISO="$(date -u -d "+30 seconds" +"%Y-%m-%dT%H:%M:%SZ" 2>/dev/null || python3 - <<'PY'\nfrom datetime import datetime, timedelta, timezone\nprint((datetime.now(timezone.utc) + timedelta(seconds=30)).strftime('%Y-%m-%dT%H:%M:%SZ'))\nPY\n)"
  printf '{"active":true,"requested_at":"%s","expires_at":"%s","mode":"%s","source":"bash_fallback"}' \
    "$NOW_ISO" "$EXPIRES_ISO" "$MODE" > "$OMC_STATE/sessions/$SESSION_ID/cancel-signal-state.json"
fi

# Clear legacy state only if no session ID (avoid clearing another session's state)
if [ -z "$SESSION_ID" ]; then
  rm -f "$OMC_STATE/${MODE}-state.json"
fi

Auto-Detection

/oh-my-claudecode:cancel follows the session-aware state contract:

By default the command inspects the current session via state_list_active and state_get_status, navigating .omc/state/sessions/{sessionId}/… to discover which mode is active.

When a session id is provided or already known, that session-scoped path is authoritative. Legacy files in .omc/state/*.json are consulted only as a compatibility fallback if the session id is missing or empty.

Swarm is a shared SQLite/marker mode (.omc/state/swarm.db / .omc/state/swarm-active.marker) and is not session-scoped.

The default cleanup flow calls state_clear with the session id to remove only the matching session files; modes stay bound to their originating session.

Active modes are still cancelled in dependency order:

Autopilot (includes linked ralph/ultraqa/ cleanup)

Ralph (cleans its linked ultrawork or )

Ultrawork (standalone)

UltraQA (standalone)

Swarm (standalone)

Ultrapilot (standalone)

Pipeline (standalone)

Team (Claude Code native)

OMC Teams (tmux CLI workers)

Plan Consensus (standalone)

Self-Improve (standalone — clear state, clean orphaned worktrees, preserve iteration_state for resume, set status: "user_stopped" in the resolved <self-improve-root>/state/agent-settings.json; new runs use .omc/self-improve/topics/<topic-slug>/, with flat .omc/self-improve/ retained only for legacy single-track resumes)

Force Clear All

Use --force or --all when you need to erase every session plus legacy artifacts, e.g., to reset the workspace entirely.

/oh-my-claudecode:cancel --force

/oh-my-claudecode:cancel --all

Steps under the hood:

state_list_active enumerates .omc/state/sessions/{sessionId}/… to find every known session.

state_clear runs once per session to drop that session’s files.

A global state_clear without session_id removes legacy files under .omc/state/*.json, .omc/state/swarm*.db, and compatibility artifacts (see list).

Team artifacts (~/.claude/teams/*/, ~/.claude/tasks/*/, .omc/state/team-state.json) are best-effort cleared as part of the legacy fallback.

Cancel for native team does NOT affect omc-teams state, and vice versa.

Every state_clear command honors the session_id argument, so even force mode still uses the session-aware paths first before deleting legacy files.

Legacy compatibility list (removed only under --force/--all):

.omc/state/autopilot-state.json

.omc/state/ralph-state.json

.omc/state/ralph-plan-state.json

.omc/state/ralph-verification.json

.omc/state/ultrawork-state.json

.omc/state/ultraqa-state.json

.omc/state/swarm.db

.omc/state/swarm.db-wal

.omc/state/swarm.db-shm

.omc/state/swarm-active.marker

.omc/state/swarm-tasks.db

.omc/state/ultrapilot-state.json

.omc/state/ultrapilot-ownership.json

.omc/state/pipeline-state.json

.omc/state/omc-teams-state.json

.omc/state/plan-consensus.json

.omc/state/ralplan-state.json

.omc/state/boulder.json

.omc/state/hud-state.json

.omc/state/subagent-tracking.json

.omc/state/subagent-tracker.lock

.omc/state/rate-limit-daemon.pid

.omc/state/rate-limit-daemon.log

.omc/state/checkpoints/ (directory)

.omc/state/sessions/ (empty directory cleanup after clearing sessions)

Implementation Steps

When you invoke this skill:

1. Parse Arguments

# Check for --force or --all flags
FORCE_MODE=false
if [[ "$*" == *"--force"* ]] || [[ "$*" == *"--all"* ]]; then
  FORCE_MODE=true
fi

2. Detect Active Modes

The skill now relies on the session-aware state contract rather than hard-coded file paths:

Call state_list_active to enumerate .omc/state/sessions/{sessionId}/… and discover every active session.

For each session id, call state_get_status to learn which mode is running (autopilot, ralph, ultrawork, etc.) and whether dependent modes exist.

If a session_id was supplied to /oh-my-claudecode:cancel, skip legacy fallback entirely and operate solely within that session path; otherwise, consult legacy files in .omc/state/*.json only if the state tools report no active session. Swarm remains a shared SQLite/marker mode outside session scoping.

Any cancellation logic in this doc mirrors the dependency order discovered via state tools (autopilot → ralph → …).

3A. Force Mode (if --force or --all)

Use force mode to clear every session plus legacy artifacts via state_clear. Direct file removal is reserved for legacy cleanup when the state tools report no active sessions.

3B. Smart Cancellation (default)

If Team Active (Claude Code native)

Teams are detected by checking for config files in ${CLAUDE_CONFIG_DIR:-~/.claude}/teams/:

# Check for active teams
TEAM_CONFIGS=$(find "${CLAUDE_CONFIG_DIR:-$HOME/.claude}"/teams -name config.json -maxdepth 2 2>/dev/null)

Two-pass cancellation protocol:

Pass 1: Graceful Shutdown

For each team found in ${CLAUDE_CONFIG_DIR:-~/.claude}/teams/:
  1. Read config.json to get team_name and members list
  2. For each non-lead member:
     a. Send shutdown_request via SendMessage
     b. Wait up to 15 seconds for shutdown_response
     c. If response received: member terminates and is auto-removed
     d. If timeout: mark member as unresponsive, continue to next
  3. Log: "Graceful pass: X/Y members responded"

Pass 2: Reconciliation

After graceful pass:
  1. Re-read config.json to check remaining members
  2. If only lead remains (or config is empty): proceed to TeamDelete
  3. If unresponsive members remain:
     a. Wait 5 more seconds (they may still be processing)
     b. Re-read config.json again
     c. If still stuck: attempt TeamDelete anyway
     d. If TeamDelete fails: report manual cleanup path

TeamDelete + Cleanup:

  1. Call TeamDelete() — removes ~/.claude/teams/{name}/ and ~/.claude/tasks/{name}/
  2. Clear team state: state_clear(mode="team")
  3. Check for linked ralph: state_read(mode="ralph") — if linked_team is true:
     a. Clear ralph state: state_clear(mode="ralph")
     b. Clear linked ultrawork if present: state_clear(mode="ultrawork")
  4. Run orphan scan (see below)
  5. Emit structured cancel report

Orphan Detection (Post-Cleanup):

After TeamDelete, verify no agent processes remain:

node "${CLAUDE_PLUGIN_ROOT}/scripts/cleanup-orphans.mjs" --team-name "{team_name}"

The orphan scanner:

Checks ps aux (Unix) or tasklist (Windows) for processes with --team-name matching the deleted team

For each orphan whose team config no longer exists: sends SIGTERM, waits 5s, sends SIGKILL if still alive

Reports cleanup results as JSON

Use --dry-run to inspect without killing. The scanner is safe to run multiple times.

Structured Cancel Report:

Team "{team_name}" cancelled:
  - Members signaled: N
  - Responses received: M
  - Unresponsive: K (list names if any)
  - TeamDelete: success/failed
  - Manual cleanup needed: yes/no
    Path: ~/.claude/teams/{name}/ and ~/.claude/tasks/{name}/

Implementation note: The cancel skill is executed by the LLM, not as a bash script. When you detect an active team:

Read ${CLAUDE_CONFIG_DIR:-~/.claude}/teams/*/config.json to find active teams

If multiple teams exist, cancel oldest first (by createdAt)

For each non-lead member, call SendMessage(type: "shutdown_request", recipient: member-name, content: "Cancelling")

Wait briefly for shutdown responses (15s per member timeout)

Re-read config.json to check for remaining members (reconciliation pass)

Call TeamDelete() to clean up

Clear team state: state_clear(mode="team", session_id)

Report structured summary to user

If Autopilot Active

Autopilot handles its own cleanup including linked ralph and ultraqa.

Read autopilot state via state_read(mode="autopilot", session_id) to get current phase

Check for linked ralph via state_read(mode="ralph", session_id):

If ralph is active and has linked_ultrawork: true, clear ultrawork first: state_clear(mode="ultrawork", session_id)

Clear ralph: state_clear(mode="ralph", session_id)

Check for linked ultraqa via state_read(mode="ultraqa", session_id):

If active, clear it: state_clear(mode="ultraqa", session_id)

Mark autopilot inactive (preserve state for resume) via state_write(mode="autopilot", session_id, state={active: false, ...existing})

If Ralph Active (but not Autopilot)

Read ralph state via state_read(mode="ralph", session_id) to check for linked ultrawork

If linked_ultrawork: true:

Read ultrawork state to verify linked_to_ralph: true

If linked, clear ultrawork: state_clear(mode="ultrawork", session_id)

Clear ralph: state_clear(mode="ralph", session_id)

If Ultrawork Active (standalone, not linked)

Read ultrawork state via state_read(mode="ultrawork", session_id)

If linked_to_ralph: true, warn user to cancel ralph instead (which cascades)

Otherwise clear: state_clear(mode="ultrawork", session_id)

If UltraQA Active (standalone)

Clear directly: state_clear(mode="ultraqa", session_id)

No Active Modes

Report: "No active OMC modes detected. Use --force to clear all state files anyway."

Implementation Notes

The cancel skill runs as follows:

Parse the --force / --all flags, tracking whether cleanup should span every session or stay scoped to the current session id.

Use state_list_active to enumerate known session ids and state_get_status to learn the active mode (autopilot, ralph, ultrawork, etc.) for each session.

When operating in default mode, call state_clear with that session_id to remove only the session’s files, then run mode-specific cleanup (autopilot → ralph → …) based on the state tool signals.

In force mode, iterate every active session, call state_clear per session, then run a global state_clear without session_id to drop legacy files (.omc/state/*.json, compatibility artifacts) and report success. Swarm remains a shared SQLite/marker mode outside session scoping.

Team artifacts (~/.claude/teams/*/, ~/.claude/tasks/*/, .omc/state/team-state.json) remain best-effort cleanup items invoked during the legacy/global pass.

Always clear skill-active state as the final step, regardless of which mode was active or whether --force was used:
state_clear(mode="skill-active", session_id)

This ensures the stop hook does not keep firing skill-protection reinforcements after cancel due to a stale skill-active-state.json. See issue #2118.

State tools always honor the session_id argument, so even force mode still clears the session-scoped paths before deleting compatibility-only legacy state.

Mode-specific subsections below describe what extra cleanup each handler performs after the state-wide operations finish.

Messages Reference

Mode
Success Message

Autopilot
"Autopilot cancelled at phase: {phase}. Progress preserved for resume."

Ralph
"Ralph cancelled. Persistent mode deactivated."

Ultrawork
"Ultrawork cancelled. Parallel execution mode deactivated."

UltraQA
"UltraQA cancelled. QA cycling workflow stopped."

Swarm
"Swarm cancelled. Coordinated agents stopped."

Ultrapilot
"Ultrapilot cancelled. Parallel autopilot workers stopped."

Pipeline
"Pipeline cancelled. Sequential agent chain stopped."

Team
"Team cancelled. Teammates shut down and cleaned up."

Plan Consensus
"Plan Consensus cancelled. Planning session ended."

Force
"All OMC modes cleared. You are free to start fresh."

None
"No active OMC modes detected."

What Gets Preserved

Mode
State Preserved
Resume Command

Autopilot
Yes (phase, files, spec, plan, verdicts)
/oh-my-claudecode:autopilot

Ralph
No
N/A

Ultrawork
No
N/A

UltraQA
No
N/A

Swarm
No
N/A

Ultrapilot
No
N/A

Pipeline
No
N/A

Plan Consensus
Yes (plan file path preserved)
N/A

Notes

Dependency-aware: Autopilot cancellation cleans up Ralph and UltraQA

Link-aware: Ralph cancellation cleans up linked Ultrawork

Safe: Only clears linked Ultrawork, preserves standalone Ultrawork

Local-only: Clears state files in .omc/state/ directory

Resume-friendly: Autopilot state is preserved for seamless resume

Team-aware: Detects native Claude Code teams and performs graceful shutdown

MCP Worker Cleanup

When cancelling modes that may have spawned MCP workers (team bridge daemons), the cancel skill should also:

Check for active MCP workers: Look for heartbeat files at .omc/state/team-bridge/{team}/*.heartbeat.json

Send shutdown signals: Write shutdown signal files for each active worker

Kill tmux sessions: Run tmux kill-session -t omc-team-{team}-{worker} for each worker

Clean up heartbeat files: Remove all heartbeat files for the team

Clean up shadow registry: Remove .omc/state/team-mcp-workers.json

Force Clear Addition

When --force is used, also clean up:

rm -rf .omc/state/team-bridge/       # Heartbeat files
rm -f .omc/state/team-mcp-workers.json  # Shadow registry
# Kill all omc-team-* tmux sessions
tmux list-sessions -F '#{session_name}' 2>/dev/null | grep '^omc-team-' | while read s; do tmux kill-session -t "$s" 2>/dev/null; done