prd
v2.88 Key Changes (MODEL-AGNOSTIC)
- Model-agnostic: Uses model configured in
~/.claude/settings.jsonor CLI/env vars - No flags required: Works with the configured default model
- Flexible: Works with GLM-5, Claude, Minimax, or any configured model
- Settings-driven: Model selection via
ANTHROPIC_DEFAULT_*_MODELenv vars
PRD System
Generate and manage Product Requirements Documents (PRDs) with INVEST-compliant user stories for structured task breakdown.
When to Use
- Planning new features or major enhancements
- Breaking down complex projects into user stories
- Creating structured task lists for Ralph Loop execution
- Documenting requirements for team collaboration
- Converting PRDs into actionable implementation tasks
PRD Structure
| Section | Purpose |
|---|---|
| Overview | Brief description and business value |
| Goals | Measurable objectives |
| User Stories | INVEST-compliant stories with acceptance criteria |
| Technical Requirements | Architecture, stack, dependencies, security |
| Success Criteria | Metrics and targets |
| Out of Scope | What's explicitly NOT included |
| Implementation Plan | Phased tasks breakdown |
| Risks & Mitigations | Potential issues and solutions |
Commands
Create PRD
ralph prd create "Implement OAuth2 authentication"
ralph prd create "Add real-time notifications" --priority high
Creates file in tasks/prd-<feature>.md
Convert to Stories
ralph prd convert tasks/prd-auth.md
Creates actionable user stories in tasks/prd-auth.json
Show Status
ralph prd status
Shows progress across all PRDs
Get Next Story
ralph prd next
Returns next uncompleted story
User Story Format (INVEST)
As a {{persona}},
I want to {{action}},
So that {{benefit}}.
Acceptance Criteria:
- [ ] {{criterion_1}}
- [ ] {{criterion_2}}
- [ ] {{criterion_3}}
INVEST Principles:
- Independent: Can be completed standalone
- Negotiable: Details can be adjusted
- Valuable: Provides user/business value
- Estimable: Complexity can be estimated
- Small: Fits within iteration limits
- Testable: Clear acceptance criteria
Integration with Ralph Loop
0. PRD CREATION → ralph prd create "feature"
1. /clarify → Intensive questions (populate PRD)
2. /classify → Complexity routing
3. PLAN → User approval (review PRD)
4. PRD CONVERSION → ralph prd convert tasks/prd-feature.md
5. EXECUTION → ralph loop --prd tasks/prd-feature.json
6. /gates → Quality validation per story
7. /retrospective → Propose PRD improvements
→ VERIFIED_DONE
Best Practices
- Start with Overview - Clear problem statement before solutions
- Measurable Goals - Use specific metrics (e.g., "reduce load time by 30%")
- INVEST Stories - Ensure each story is independent and testable
- Scope Management - Explicitly document "Out of Scope" items
- Risk Assessment - Identify risks early with mitigations
- Stakeholder Review - Get approval before converting to stories
- Iterative Execution - Execute one story at a time with validation
Related Skills
/clarify- Intensive questions (populate PRD details)/iterate- Execute PRD stories iteratively/orchestrator- Full workflow with PRD integration/plan- Plan-state management for execution tracking
More from alfredolopez80/multi-agent-ralph-loop
stop-slop
A skill for removing AI-generated writing patterns ('slop') from prose. Eliminates telltale signs of AI writing like filler phrases, excessive hedging, overly formal language, and mechanical sentence structures. Use when: writing content that should sound human and natural, editing AI-generated drafts, cleaning up prose for publication, or any content that needs to sound authentic rather than AI-generated. Triggers: 'stop-slop', 'remove AI tells', 'clean up prose', 'make it sound human', 'edit AI writing'.
10iterate
Ralph Loop pattern with swarm mode: iterative execution until VERIFIED_DONE with multi-agent coordination. Use when: (1) iterative refinement needed, (2) quality gates must pass, (3) automated validation required. Triggers: /iterate, 'iterate until done', 'keep trying', 'fix until passing', 'loop until done'.
2security
Security audit with Codex + MiniMax second opinion. Integrates ralph-security agent (6 quality pillars, OWASP A01-A10). Uses LSP for code navigation during analysis. Use when: (1) /security is invoked, (2) task relates to security functionality.
1adr
Architecture Decision Records management. Actions: create (new ADR), list (show all), search (find by keyword). Use when: (1) making architecture decisions, (2) choosing between technologies, (3) documenting trade-offs. Triggers: /adr, 'architecture decision', 'decision record', 'document decision'.
1curator
Full curator pipeline for autonomous learning from quality repositories. Executes: discovery → scoring → ranking → ingest → learn → vault sync. Writes to procedural memory AND Obsidian vault for Graph View visualization and graduation pipeline. Use for: populating procedural memory with domain patterns, first-time domain learning, comprehensive knowledge building. Triggers: /curator full, 'learn patterns from repos', 'build knowledge base'.
1parallel
Run multiple Ralph loops concurrently for independent tasks. Supports all 6 ralph-* teammates (coder, reviewer, tester, researcher, frontend, security). Manages parallel agent execution with proper isolation and result aggregation. Use when: (1) multiple independent fixes needed, (2) parallel reviews required, (3) batch processing tasks. Triggers: /parallel, 'parallel loops', 'concurrent execution', 'run in parallel', 'batch'.
1