Rpi
- 6 repo stars
- Updated July 18, 2026
- A-NGJ/rpi
rpi is a Claude Code plugin and Go MCP CLI that turns feature requests into staged, reviewable artifacts and phased implementation so developers keep control of AI coding work.
About
rpi is a Claude Code plugin implementing the Research-Propose-Plan-Implement flow for steering capable coding agents with reviewable artifacts at every stage. A compiled Go CLI exposed over MCP handles scaffolding, frontmatter, chains, and status so the LLM spends tokens on reasoning, not file bookkeeping. Use it when you want explicit checkpoints, durable .rpi specs and plans, and phase-by-phase implementation with verification before shipping features.
- Staged /rpi-research through /rpi-verify artifact pipeline
- Go MCP server offloads parsing, chains, and status to binary
- Human approval gates between propose, plan, and implement phases
- Persistent .rpi specs, plans, and reviews across sessions
Rpi by the numbers
- Data as of Jul 19, 2026 (Skillselion catalog sync)
/plugin marketplace add A-NGJ/rpiAdd your badge
Show developers this marketplace is listed on Skillselion. Paste this into your README.
| repo stars | ★ 6 |
|---|---|
| Last updated | July 18, 2026 |
| Repository | A-NGJ/rpi ↗ |
How do I steer an AI coding agent through research, design, planning, and implementation with reviewable checkpoints instead of one-shot prompts?
Drive Research-Propose-Plan-Implement workflows with reviewable .rpi artifacts and a Go MCP CLI that saves tokens.
Who is it for?
Developers and teams who want human-in-the-loop staged workflows, durable .rpi artifacts, and token-efficient MCP bookkeeping.
Skip if: Quick one-off edits where a single unstructured prompt is enough and no cross-session artifact trail is needed.
What you get
Versioned .rpi research, designs, specs, plans, reviews, commits, and optional diagnoses with session resume via rpi status.
Plugins in this marketplace
1 plugin - install individually after you add the marketplace.
Recommended Marketplaces
FAQ
What artifacts does RPI create?
Files under .Rpi/ including research notes, designs, behavioral specs, phased plans, reviews, and diagnoses with inherited decision provenance.
Why use the Go MCP server?
Rpi serve exposes typed tools so parsing, chain traversal, checkbox counts, and status run outside the LLM context, saving tokens.
When should I skip stages?
Small bug fixes can go Plan → Implement; medium features use Propose → Plan → Implement; large work keeps the full staged path.