5-stage quality pipeline for Agent Teams — Plan, PRD, Execute, Verify, Fix
Scanned 5/28/2026
Install via CLI
openskills install speson/not-my-reforge---
name: team-pipeline
description: 5-stage quality pipeline for Agent Teams — Plan, PRD, Execute, Verify, Fix
user_invocable: true
---
# Team Pipeline
A structured 5-stage pipeline for executing complex tasks with Agent Teams.
## Pipeline Stages
### Stage 1: Plan
The **planner** agent creates an implementation plan:
- Analyze the codebase and task requirements
- Identify change points and dependencies
- Produce a structured plan with phases and steps
- Enter plan mode for lead approval
### Stage 2: PRD (Product Requirements Document)
The **lead** converts the approved plan into concrete tasks:
- Create TaskCreate entries for each work unit
- Define clear acceptance criteria per task
- Set task dependencies (blockedBy/blocks)
- Assign tasks to workers
### Stage 3: Execute
**Workers** implement in parallel:
- Each worker operates in an isolated git worktree
- Workers follow their task's acceptance criteria
- **Researcher** gathers information as needed (background)
- **Tester** writes tests for implementations
### Stage 4: Verify
**Reviewer** validates all changes:
- Review each worker's diff against acceptance criteria
- Run full test suite
- Check for type errors and lint issues
- Verdict: APPROVE or REQUEST_CHANGES
### Stage 5: Fix
If verification fails:
- Failed items return to Execute stage
- Workers address reviewer feedback
- Re-verify until all checks pass
- Maximum 3 Fix cycles before escalating to user
## Usage
```
omo team 3 <task description>
```
The pipeline runs automatically when using the team command with the team-pipeline skill.
## Quality Gates
Each stage transition requires:
- Plan → PRD: Lead approval of plan
- PRD → Execute: All tasks created with acceptance criteria
- Execute → Verify: All tasks marked complete
- Verify → Fix/Done: Reviewer verdict
- Fix → Verify: Worker fixes applied
No comments yet. Be the first to comment!
Ultra-compressed communication mode. Cuts token usage ~75% by speaking like caveman while keeping full technical accuracy. Supports intensity levels: lite, full (default), ultra, wenyan-lite, wenyan-full, wenyan-ultra. Use when user says "caveman mode", "talk like caveman", "use caveman", "less tokens", "be brief", or invokes /caveman. Also auto-triggers when token efficiency is requested.