Add comprehensive execution protocol to formula description:
- EXECUTION LOOP: find ready step, close it, repeat
- ON RESUME: same as fresh start, molecule IS the state
- DO/DO NOT guidelines for autonomous execution
- MONITORING guidance for mega-molecules
This enables agents to execute Hanoi autonomously through restarts,
proving nondeterministic idempotence at scale.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add Step 3a to survey-workers that automatically nukes idle polecats
with clean git state. Only escalates to Mayor when uncommitted or
unpushed work is detected.
This reduces mail noise - idle polecats without work are pure overhead
and can be safely cleaned up without human intervention.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Changed table entry from 'Skip (no action needed)' to 'Auto-nuke if clean (Step 3a)'.
TODO: Add Step 3a section with full auto-nuke logic:
- Check git status for uncommitted changes
- Check for unpushed commits
- If clean: gt polecat nuke <name>
- If dirty: escalate to Mayor for recovery decision
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Adds explicit priming commands at the start of workspace assessment:
- gt prime: Load gt/mol context
- bd prime: Load beads context
Critical for fresh sessions, post-compaction, or context clear scenarios.
Also updates exit criteria to reflect priming requirement.
(gt-0xhuz)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add configurable build_command and test_command variables to the
mol-sync-workspace formula. Previously hardcoded to Go commands,
now supports any project type via variable configuration.
- Add build_command variable (default: go build ./...)
- Add test_command variable (default: go test ./...)
- Update verify-build step to use {{ build_command }}
- Update run-tests step to use {{ test_command }}
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Role templates (crew, polecat, mayor, deacon, witness, refinery)
- prime.go startup protocol messages
- Documentation (propulsion, reference, molecules, wisp architecture)
- Session hints and sling prompts
- Formula template instructions
- CLAUDE.md
The hook is the user-facing concept, molecules are implementation details.
Agents should use `gt hook` to check what work is assigned to them.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The health-scan step now includes gt nudge commands that serve dual purposes:
1. Liveness verification - agent responds to prove it's alive
2. Backoff reset - any nudge resets agent's backoff to base interval
This ensures patrol agents (witness/refinery) remain responsive even during
quiet periods when the feed has no mutations. Since Deacon patrols every
~1-2 minutes, maximum backoff is bounded by the ping interval.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Molecule for coordinated workspace sync across agents:
- Git pull/rebase with conflict preservation
- Beads sync and doctor check
- Build/test verification
- Worktree cleanup
- Readiness reporting
Designed for broadcast to all agents before batch assignments.
Boot is a watchdog that the daemon pokes instead of Deacon directly,
centralizing the 'when to wake Deacon' decision in an agent that can
reason about context.
Key changes:
- Add internal/boot package with marker file and status tracking
- Add gt boot commands: status, spawn, triage
- Add mol-boot-triage formula for Boot's triage cycle
- Modify daemon to call ensureBootRunning instead of ensureDeaconRunning
- Add tmux.IsAvailable() for degraded mode detection
- Add boot.md.tmpl role template
Boot lifecycle:
1. Daemon tick spawns Boot (fresh each time)
2. Boot runs triage: observe, decide, act
3. Boot cleans stale handoffs from Deacon inbox
4. Boot exits (or handoffs in non-degraded mode)
In degraded mode (no tmux), Boot runs mechanical triage directly.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Major changes to mol-deacon-patrol.formula.toml v4:
- Add DOG_DONE message handling in inbox-check step
- Add dog-pool-maintenance step for idle dog availability
- Change orphan-check to detect-only (dispatches to dogs via sling)
- Change session-gc to detect-only (dispatches to dogs via sling)
The Deacon now stays lightweight by detecting issues and dispatching
heavy work to dogs rather than executing inline. This prevents context
bloat and keeps patrol cycles fast.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Create five formulas for dog-executed infrastructure tasks:
- mol-convoy-cleanup: Archive convoy, notify overseer
- mol-dep-propagate: Cross-rig dependency resolution
- mol-digest-generate: Daily digest for overseer
- mol-orphan-scan: Find and reassign orphaned work
- mol-session-gc: Clean stale sessions
Each formula follows the standard TOML structure with:
- Description and dog contract
- Step-by-step workflow with dependencies
- Variable definitions
- Squash configuration for digests
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Added three new patrol steps for convoy orchestration:
- check-convoy-completion: Watch for tracked issue closures, auto-close convoys
- resolve-external-deps: Propagate cross-rig dependency resolution
- fire-notifications: Notify Overseer/Witnesses of convoy and dep events
Updated health-scan to depend on fire-notifications (runs after convoy steps).
(gt-wthcc)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Adds a new workflow formula for polecat work that includes strategic
context reading before design. The context_docs variable allows
customization of which organizational documents to read.
Default: ~/gt/docs/hop/CONTEXT.md,~/gt/docs/PRIMING.md (Gas Town docs)
Override: gt sling formula --var context_docs="docs/ARCH.md,docs/STYLE.md"
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Update mol-witness-patrol formula to:
1. Close the epic (swarm molecule) when all children are closed
2. Close the swarm tracking wisp
3. Notify Mayor with SWARM_COMPLETE mail
Activity feed events are generated automatically by bd close commands.
(gt-kc7yj.3)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
When ready tasks exist but no idle polecats are available, the Witness
should log the situation rather than silently skip.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Add dispatch-swarm-work step to mol-witness-patrol formula that:
- Queries bd ready --parent=<epic> to find ready swarm tasks
- Finds idle polecats via agent beads
- Dispatches work via gt sling <task> <rig>/<polecat>
Also updated:
- SWARM mail handling to store epic_id in labels
- check-swarm-completion to use beads discovery model
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Notifications now come immediately after push, before cleanup
- Added STOP HERE warning after push
- Made verification gate more emphatic
- Cleanup happens only after notifications confirmed
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Updated mol-refinery-patrol.formula.toml:
- Added MR bead ID tracking to inbox-check MERGE_READY parsing
- Restructured merge-push into 3 clear steps with REQUIRED callouts
- Added Step 2: MERGED notification (DO NOT SKIP)
- Added Step 3: Close MR bead (DO NOT SKIP)
- Added verification checklist before proceeding
- Added generate-summary verification to catch missed notifications
This ensures refinery agents complete the full lifecycle and trigger
witness cleanup (gt polecat nuke) after merge.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Complete redesign of polecat work formula with 8 concrete steps:
1. load-context - Prime, read issue, verify can proceed
2. branch-setup - Clean feature branch, rebased on main
3. implement - Do work, commit regularly, file discovered work
4. self-review - Review diff, fix issues before testing
5. run-tests - ALL tests must pass, verify coverage
6. cleanup-workspace - No uncommitted, no stash, no cruft
7. close-issue - bd close with summary
8. signal-complete - POLECAT_DONE mail, WAIT for termination
Also updated mol-polecat-lease.formula.toml (v2):
- 5 steps: boot → working → verifying → merge_requested → done
- Aligns with witness patrol and mail protocol
- Clear verification criteria before MERGE_READY
Added templates/polecat-CLAUDE.md:
- Polecat role context for spawned workers
- Propulsion principle, key commands, completion protocol
- Clear "do NOT" section (don't exit, don't push to main)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Added WITNESS_PING protocol for monitoring Deacon health:
Witness patrol (mol-witness-patrol):
- Added ping-deacon step after survey-workers
- Sends WITNESS_PING mail to Deacon each patrol cycle
- Checks Deacon agent bead last_activity timestamp
- Escalates to Mayor if Deacon appears unresponsive
Deacon patrol (mol-deacon-patrol):
- Added WITNESS_PING handling in inbox-check
- Added second-order monitoring section to description
- Bumped formula version to 2
This prevents the "who watches the watchers" problem - if Deacon dies,
the collective Witness fleet detects it and escalates.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Updated mol-refinery-patrol formula with:
- MERGE_READY handling in inbox-check (parse branch/issue/polecat)
- Mail-based queue instead of git branch scanning
- MERGED mail sent to Witness after successful merge
- Flow diagram showing Witness→Refinery→Witness mail protocol
- Bumped formula version to 2
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Changes:
1. prime.go: Create agent beads on first prime for all roles including crew
2. done.go: Update agent state to done/stuck/idle when work completes
3. sling.go: Populate hook_bead when work is assigned to agents
4. mol-witness-patrol.formula.toml: Document dead and spawning states
Closes: gt-hymm0, gt-0lop3, gt-59k2x, gt-c4j4j
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Updated mol-witness-patrol formula with:
- MERGE_READY mail sent after polecat verification (process-cleanups)
- MERGED mail handling in inbox-check (completes cleanup)
- Cleanup wisp state machine: pending → merge-requested → closed
- Fallback discovery in survey-workers creates cleanup if done polecat
found without existing wisp
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Changes:
1. Fix reportAgentState in prime.go to use beads API directly instead of
non-existent `bd agent state` command. Agents now properly self-report
their state to their agent beads on startup.
2. Update witness patrol survey-workers step to use agent beads:
- List polecats via `bd list --type=agent --json`
- Filter by role_type: polecat in description
- Check agent_state field (running/idle/stuck/done)
- Trust agent-reported state (ZFC principle)
No more PID/tmux inference for polecat state - agents self-report.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Beads CLI restructured: pour and wisp are now subcommands of mol.
- bd pour → bd mol pour
- bd wisp → bd mol wisp
Updated all documentation, templates, and code to use new command structure.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Delete all .formula.json files (TOML versions exist)
- Update sling.go comment to say TOML only
Note: Doc updates were prepared but those docs were deleted
upstream in the recent refactoring.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
New step evaluates pending async gates each patrol cycle:
- Lists open gates via bd gate list
- For timer gates, checks if elapsed time exceeds timeout
- Closes elapsed gates and notifies waiters
- Runs in parallel with trigger-pending-spawns
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Converted all .formula.json files to .formula.toml using bd formula convert.
TOML provides better ergonomics:
- Multi-line strings without \n escaping
- Human-readable diffs
- Comments allowed
Original JSON files retained for backwards compatibility.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Implement WispGCCheck in internal/doctor/wisp_check.go
- Scans rigs for wisps older than 1 hour threshold
- Fix runs `bd --no-daemon wisp gc` in each affected rig
- Register wisp-gc check in gt doctor
- Update help text to document cleanup checks
- Simplify Deacon patrol session-gc step to just use gt doctor --fix
Now `gt doctor --fix` handles all cleanup:
- orphan-sessions: Kill orphaned tmux sessions
- orphan-processes: Kill orphaned Claude processes
- wisp-gc: Garbage collect abandoned wisps (>1h)
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The previous commit only fixed wisp gc, but the session cleanup
was also using non-existent commands. Now session-gc step properly
documents:
1. gt doctor --fix for orphaned sessions and processes
2. bd --no-daemon wisp gc for orphaned wisps
Also noted that log pruning and molecule archival are handled
elsewhere (not implemented / bd mol squash|burn respectively).
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The session-gc step was referencing non-existent `gt gc --wisps` command.
Updated to use the correct `bd --no-daemon wisp gc` command with examples.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
Convert mol-polecat-lease and mol-polecat-work formulas from the
old "variables" array format to the canonical "vars" map format
used by other formulas.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Fix security-audit.formula.json to match Go schema (advice as array)
- Update pointcuts and targets for exact matching
- Add shiny-secure.formula.json demonstrating aspect composition
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
The formula used "variables" as an array, but the schema expects
"vars" as a map. This prevented variable metadata (required, default)
from being parsed by bd cook.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
- Add required title fields to shiny.formula.json steps
- Add required title fields to rule-of-five.formula.json template steps
- Fix shiny-enterprise.formula.json schema: use compose.expand array format
- Add type: "workflow" to shiny.formula.json
These changes align formulas with the expected schema in bd cook.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
