AI-Optimized PRD Writing
Create PRDs that AI coding assistants can actually execute effectively.
New: /octo:prd Command
/octo:prd user authentication feature
/octo:prd checkout flow redesignFeatures:
- Multi-phase PRD generation workflow
- Self-scoring against 100-point framework
- Templates for lightweight, standard, and comprehensive PRDs
- Integration with research and debate workflows
Enhanced: product-writer Persona
Completely rewritten with AI-specific PRD patterns:
| Pattern | Before | After |
|---|---|---|
| Structure | Holistic features | Sequential, dependency-ordered phases |
| Requirements | Generic | FR codes with P0/P1/P2 priorities |
| Boundaries | Implied | Explicit Non-Goals section |
| Work sizing | Undefined | 5-15 min phases for frontier LLMs |
| Acceptance | Vague | Given-When-Then testable criteria |
| Scoring | None | 100-point self-validation |
PRD Scoring Framework (100 Points)
Based on 2026 AI coding assistant research:
- AI-Specific Optimization (25 pts): Sequential phases, explicit boundaries, structured format
- Traditional PRD Core (25 pts): Problem, goals, personas, technical specs
- Implementation Clarity (30 pts): Functional/non-functional requirements, architecture, phases
- Completeness (20 pts): Risks, dependencies, examples, documentation quality
Score Interpretation:
- 90-100: Excellent - Ready for AI implementation
- 80-89: Good - Minor gaps
- 70-79: Acceptable - Needs optimization
- <70: Needs revision
Why This Matters
Traditional PRDs fail with AI because they're written for humans who infer context. AI coding assistants need:
- Sequential, dependency-ordered phases (not holistic feature descriptions)
- Explicit boundaries (AI cannot infer from omission)
- Testable acceptance criteria (not vague success definitions)
- Right-sized work units (5-15 minutes per phase)
Full Changelog: https://github.com/nyldn/claude-octopus/blob/main/CHANGELOG.md