mission-driver
Create roadmaps and mission configs, then drive the mission-driver AI dev-loop engine. mission-driver lives at tools/mission-driver/ — it reads missions/<name>.json and loops CHECK → REVIEW_PLANS → EXEC_PLANS → DRAFT_PLANS → DEEP_AUDIT until the roadmap is done or the audit budget is exhausted. Use this skill when the user wants to: (1) plan a development goal as a roadmap ("create roadmap", "make a roadmap", "new dev goal"); (2) bootstrap a mission config from an existing roadmap ("create mission", "config mission", "draft mission"); (3) launch / resume / monitor a mission ("run mission", "st
Profile →npx skills add entropy-cloud/attractor-guided-engineering-template --skill mission-driver --agent claude-code
Same command for any agent — swap --agent for codex, cursor, copilot.
Weekly change comes from our own snapshots, not the repository page — it measures attention, not adoption.
What it does
The skill operates an engine at tools/mission-driver/ that reads missions/<name>.json and runs a loop consisting of CHECK → REVIEW_PLANS → EXEC_PLANS → DRAFT_PLANS → DEEP_AUDIT until either the roadmap is complete or the audit budget is exhausted. It supports creating roadmaps from a goal description, bootstrapping mission configs from roadmaps, and launching/resuming/monitoring missions. Roadmaps are authored in docs/backlog and mission configs in missions/<name>.json with associated directories for plans and audits. It logs runtime state to _tmp/** and uses opencode run processes for each AI step. It enforces a fixed plan lifecycle: draft → REVIEW_PLANS → active → EXEC_PLANS → completed, with deep audits if needed.
How it works
- It reads a fixed config structure (missions/<name>.json) and uses runtime state in _tmp/<runDir>/run-state.json.
- Each AI step runs as a child opencode run process; logs land in _tmp/<ts>-mission-driver/.
- The loop sequence is CHECK, REVIEW_PLANS (promotes drafts to active), EXEC_PLANS (per active plan: EXECUTE → CLOSURE_SCRIPT_CHECK → CLOSURE_AUDIT → BUILD_VERIFY), DRAFT_PLANS (drafts 1-3 plans from roadmap and reviews them), then repeats until nothing to draft or maxAuditRounds is reached, at which point DEEP_AUDIT can occur.
- The draft command can generate both a roadmap and missions/<name>.json from a single description, after scope brief is approved. Mission configs require explicit commands to be defined, even when extending base defaults.
- Plan format is enforced by tools/mission-driver/src/plan-check.mjs; command sets are run via stack-specific test/build/typecheck commands.
When to use it
- Triggered when a development goal needs a roadmap, or a mission config must be created from a roadmap, or a mission should be run/resumed/monitored.
- Use for roadmaps that precede a mission and for multi-step work that requires audits and verifications.
What it can touch
- missions/<name>.json (the mission config) and related directories: docs/backlog, docs/plans/{mission-name}, docs/audits/{mission-name}, _tmp, and logs under _tmp/**.
- The engine runs as shell-based opencode run processes; logs land in _tmp and per-step agent logs under _tmp/<runDir>/oc-<STEP>-*.log.
Caveats
- Plan lifecycle and state are managed within a fixed contract: draft → REVIEW_PLANS → active → EXEC_PLANS → completed, with DEEP_AUDIT as needed.
- The runtime state is not stored in the mission file but in _tmp/<runDir>/run-state.json.
- The tool requires proper mission config blocks (commands and tests) to be present; placeholders are not allowed for commands.test.
# mission-driver — Roadmap creation and mission execution This skill operates the `tools/mission-driver/` engine shipped with this repo. It is **stack-agnostic**: it works for Node, Python, Java, Go, or any project whose verification commands can be expressed as shell strings. Project-specific values (test command, module layout, commit format) are read from `missions/base.json`, `docs/context/project-context.md`, and `AGENTS.md` rather than hardcoded here. **Companion docs** (read on demand, not all at once): - `tools/mission-driver/docs/user-manual.zh.md` / `user-manual.en.md` — full training handbook - `tools/mission-driver/README.md` — command cheat-sheet - `tools/mission-driver/EXECUTION-PRINCIPLE.md` — internal execution deep dive - `tools/mission-driver/TROUBLESHOOTING.md` — diagnostics when stuck - `docs/plans/00-plan-authoring-and-execution-guide.md` — plan format and lifecycle - `references/mission-config-schema.md` — full mission.json schema (companion to this file) - `references/roadmap-template.md` — roadmap structure with annotated example ## Mental model A **mission** = a fixed config (`missions/<name>.json`) for one development goal. The engine reads it and enters a
- Mental model
- Decision: is mission-driver the right tool?
- Workflow A: Create a Roadmap
- A.1 Pre-flight
- A.2 Write the roadmap
- A.3 Self-check
- Workflow B: Create the mission config
- B.1 Two creation paths
- B.2 Decide field values (stack-agnostic)
- B.3 Validate the mission config
- B.4 Pre-flight: plans directory and plan-guide
- Workflow C: Run and monitor the mission
- C.1 Pre-flight checks
- C.2 Launch commands
node tools/mission-driver/src/mission-check.mjs missions/<name>.json . node tools/mission-driver/src/reap-orphans.mjs --startup _tmp <PID> Standard run (foreground, starts monitor on port 9300) Equivalent main-command form (no "run" keyword — same effect) List all available missions List a mission's steps Single-step debug (run only CHECK then exit) Resume from a specific step (run that step then continue the loop normally) Dry-run (mock agent, no real model calls — verify flow orchestration) Cap loops to prevent runaway
What does the mission-driver skill do?
Create roadmaps and mission configs, then drive the mission-driver AI dev-loop engine. mission-driver lives at tools/mission-driver/ — it reads missions/<name>.json and loops CHECK → REVIEW_PLANS → EXEC_PLANS → DRAFT_PLANS → DEEP_AUDIT until the roadmap is done or the audit budget is exhausted. Use this skill when the user wants to: (1) plan a development goal as a roadmap ("create roadmap", "make a roadmap", "new dev goal"); (2) bootstrap a mission config from an existing roadmap ("create mission", "config mission", "draft mission"); (3) launch / resume / monitor a mission ("run mission", "st
How do I install it?
Run `npx skills add entropy-cloud/attractor-guided-engineering-template --skill mission-driver --agent claude-code` — it drops the skill into your project so the agent can pick it up. Swap the --agent value for codex, cursor or copilot if you use one of those.
Where does this skill come from?
From entropy-cloud/attractor-guided-engineering-template, a repository with 67 stars. We read it straight from the repository tree rather than a submitted listing, so what you see here is what is actually published.
Is a popular skill a good skill?
Not necessarily. Stars measure attention, not adoption — a repository can trend for a week and be abandoned. That is why we show the weekly change from our own snapshots next to the total, instead of a single flattering number.