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
Explores a codebase and produces a structured spec plus a phased implementation plan with file-level task breakdowns. Takes a task description, researches the code autonomously, and writes spec.md and plan.md into a docs/plan/ directory. Does not ask follow-up questions.
Derives plans from actual codebase research rather than generic templates, so every task references real file paths and existing patterns.
When to use it
- Planning a new feature before writing any code
- Mapping which files a bug fix will touch
- Creating a phased refactor roadmap across multiple modules
- Turning a vague 'battle plan' request into 5-15 concrete tasks
- Generating a migration plan for a dependency upgrade
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: plan a user authentication feature
- 1Parse task description and auto-classify as 'feature' type
- 2Read CLAUDE.md, package.json, and docs/prd.md for project context
- 3Search codebase with Glob and Grep for existing auth-related files and tests
- 4Detect deploy infrastructure by scanning for Dockerfile, fly.toml, and similar configs
- 5Write spec.md with acceptance criteria and plan.md with 2-4 phases totaling 5-15 tasks referencing specific file paths
docs/plan/user-auth_20260313/spec.md and plan.md ready for /build