Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
Universalize: remove project-specific references, add SearXNG recommendation
What This Skill Does
Executes tasks from an implementation plan created by /plan. Reads plan.md, picks the next unchecked task, implements it using a TDD workflow, commits with conventional format, and records a SHA annotation in plan.md for traceability. Supports phase gates that verify tests and linting before moving forward.
Each task gets its own atomic commit with a SHA recorded in plan.md, so any task or entire phase can be reverted with a single git revert without touching unrelated work.
When to use it
- Working through a backlog of planned tasks one commit at a time
- Enforcing red-green-refactor TDD on each individual feature task
- Resuming an interrupted build session from the last in-progress task
- Annotating plan.md with commit SHAs for surgical per-task reverts
- Running phase-level verification checkpoints before advancing to the next phase
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: 'build it' after /plan has created a track
- 1Reads docs/plan/{trackId}/plan.md and spec.md to load the task list and acceptance criteria
- 2Finds the first unchecked task and marks it in-progress ([~]) in plan.md
- 3Writes a failing test for the task functionality (TDD red phase), then implements minimum code to pass (green phase)
- 4Runs linter, formatter, and type-checker; fixes any errors before committing
- 5Commits with a conventional commit message, captures the SHA, and appends it inline to the task entry in plan.md
Task marked [x] with SHA annotation; if all tasks in the phase are done, runs full verification suite and records a phase checkpoint SHA