majiayu000 RPI Workflow
You have access to workflow skills for structured development.
The RPI Workflow
Research → Plan → Implement → Validate
↑ │
└──── Knowledge Flywheel ────┘
Research Phase
/research <topic> # Deep codebase exploration
/knowledge <query> # Query existing knowledge
Output: .agents/research/<topic>.md
Plan Phase
/pre-mortem <spec> # Simulate failures before implementing
/plan <goal> # Decompose into trackable issues
Output: Beads issues with dependencies
Implement Phase
/implement <issue> # Single issue execution
/crank <epic> # Autonomous epic loop (uses swarm for waves)
/swarm # Parallel execution (fresh context per agent)
Output: Code changes, tests, documentation
Validate Phase
/vibe [target] # Code validation (security, quality, architecture)
/post-mortem # Extract learnings after completion
/retro # Quick retrospective
Output: .agents/learnings/, .agents/patterns/
Release Phase
/release [version] # Full release: changelog + bump + commit + tag
/release --check # Readiness validation only (GO/NO-GO)
/release --dry-run # Preview without writing
Output: Updated CHANGELOG.md, version bumps, git tag, .agents/releases/
Phase-to-Skill Mapping
| Phase | Primary Skill | Supporting Skills |
|---|---|---|
| Research | /research |
/knowledge, /inject |
| Plan | /plan |
/pre-mortem |
| Implement | /implement |
/crank (epic loop), /swarm (parallel execution) |
| Validate | /vibe |
/retro, /post-mortem |
| Release | /release |
— |
Choosing the skill:
- Use
/implementfor single issue execution. - Use
/crankfor autonomous epic execution (loops waves via swarm until done). - Use
/swarmdirectly for parallel execution without beads (TaskList only). - Use
/ratchetto gate/record progress through RPI.
Available Skills (43 user-facing)
Core Skills (start here)
| Skill | Purpose |
|---|---|
/research |
Deep codebase exploration |
/brainstorm |
Structured idea exploration before planning |
/plan |
Epic decomposition into issues |
/implement |
Execute single issue |
/vibe |
Code validation (complexity + multi-model council) |
/status |
Single-screen dashboard of current work and suggested next action |
Power Skills (when you're ready)
| Skill | Purpose |
|---|---|
/council |
Multi-model consensus review (validate, brainstorm, research) |
/pre-mortem |
Failure simulation before implementing |
/post-mortem |
Full validation + knowledge extraction |
/bug-hunt |
Root cause analysis |
/release |
Pre-flight, changelog, version bumps, tag |
/crank |
Autonomous epic loop (uses swarm for each wave) |
/doc |
Documentation generation |
/retro |
Extract learnings from completed work |
/knowledge |
Query knowledge artifacts |
/learn |
Capture knowledge manually into the flywheel |
Expert Skills (advanced workflows)
| Skill | Purpose |
|---|---|
/swarm |
Fresh-context parallel execution (Ralph pattern) |
/rpi |
Full RPI lifecycle orchestrator (research → plan → implement → validate) |
/evolve |
Goal-driven fitness-scored improvement loop |
/codex-team |
Parallel Codex agent execution |
/openai-docs |
Official OpenAI docs lookup with citations |
/oss-docs |
OSS documentation scaffold and audit |
/pr-research |
Upstream repository research before contribution |
/pr-plan |
External contribution planning |
/pr-implement |
Fork-based PR implementation |
/pr-validate |
PR-specific validation and isolation checks |
/pr-prep |
PR preparation and structured body generation |
/pr-retro |
Learn from PR outcomes |
/complexity |
Code complexity analysis |
/product |
Interactive PRODUCT.md generation |
/handoff |
Session handoff for continuation |
/inbox |
Agent mail monitoring |
/recover |
Post-compaction context recovery |
/trace |
Trace design decisions through history |
/provenance |
Trace artifact lineage to sources |
/beads |
Issue tracking operations |
/heal-skill |
Detect and fix skill hygiene issues |
/converter |
Convert skills to Codex/Cursor formats |
/update |
Reinstall all AgentOps skills from latest source |
Knowledge Flywheel
Every /post-mortem feeds back to /research:
- Learnings extracted →
.agents/learnings/ - Patterns discovered →
.agents/patterns/ - Research enriched → Future sessions benefit
Issue Tracking
This workflow uses beads for git-native issue tracking:
bd ready # Unblocked issues
bd show <id> # Issue details
bd close <id> # Close issue
bd sync # Sync with git
Examples
SessionStart Auto-Injection
Hook triggers: session-start.sh runs at session start
What happens:
- Hook injects this skill automatically into session context
- Agent loads RPI workflow overview, phase-to-skill mapping, trigger patterns
- Agent understands available skills without user prompting
- User says "check my code" → agent recognizes
/vibetrigger naturally - Agent executes skill using workflow knowledge from this reference
Result: Agent knows the full skill catalog and workflow from session start, enabling natural language skill invocation.
Workflow Reference During Planning
User says: "How should I approach this feature?"
What happens:
- Agent references this skill's RPI workflow section
- Agent recommends Research → Plan → Implement → Validate phases
- Agent suggests
/researchfor codebase exploration,/planfor decomposition - Agent explains
/pre-mortemfor failure simulation before implementation - User follows recommended workflow with agent guidance
Result: Agent provides structured workflow guidance based on this meta-skill, avoiding ad-hoc approaches.
Troubleshooting
| Problem | Cause | Solution |
|---|---|---|
| Skill not auto-loaded | Hook not configured or SessionStart disabled | Verify hooks/session-start.sh exists; check hook enable flags |
| Outdated skill catalog | This file not synced with actual skills/ directory | Update skill list in this file after adding/removing skills |
| Wrong skill suggested | Natural language trigger ambiguous | User explicitly calls skill with /skill-name syntax |
| Workflow unclear | RPI phases not well-documented here | Read full workflow guide in README.md or docs/ARCHITECTURE.md |