Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
Version History
Make launcher scripts optional, direct skill invocation as primary path
What This Skill Does
Chains multiple Claude skills into automated, sequential pipelines. Two built-in pipelines are available: Research (research → validate, producing a PRD) and Dev (scaffold → setup → plan → build, producing working code). Pipeline state is persisted to a local file so runs can be resumed if interrupted.
Uses a Stop hook to advance pipeline stages automatically on session exit, removing the need to manually invoke each skill in sequence.
When to use it
- Running a full idea-to-PRD workflow with one command
- Scaffolding and building a new project end-to-end without manual skill invocation
- Resuming a partially completed pipeline after an interrupted session
- Monitoring multi-stage automation progress via a live tmux dashboard
- Skipping already-completed stages when re-running a pipeline
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: /pipeline research "AI therapist app"
- 1Parses arguments: pipeline type = research, idea = 'AI therapist app'
- 2Confirms stages with user: /research → /validate, asks to continue
- 3Runs /research 'AI therapist app', writes research.md
- 4Stop hook detects stage completion and advances to next stage
- 5Runs /validate, writes prd.md, outputs <solo:done/> to signal completion
research.md and prd.md written to the project directory