/pair-process-specify-prd — PRD Creation

Create a comprehensive Product Requirements Document through collaborative analysis and structured information gathering. Follows a template-first approach: analyze PRD template, gather information via hypothesis-driven questions, create PRD, review and approve.

Process Profile

<!-- process-step: id=specify-prd -->

Executable form of the specify-prd step. A direct invocation while a step is disabled by the project's profile warns and asks for confirmation; a composed one never prompts — it degrades exactly as a step that is not installed. No section ⇒ no-op. See process-profile gate.

Arguments

Argument	Required	Description
$section	No	Specific PRD section to update (e.g., goals, features). If omitted, full PRD process

Phase 0: Detect Existing PRD (Idempotency Gate)

Step 0.1: Check Current State

1. Check: Does adoption/product/PRD.md exist?

2. Act (file exists): Read the file. Determine if it is a template (contains [Product/feature name] or [Creation date]) or a populated PRD.

   • If template: treat as new PRD. Proceed to Phase 1.

   • If populated and $section provided: jump to Phase 3 (selective update of that section).

   • If populated and no $section: present current PRD summary and ask:

     PRD already exists for [product name]. Options:

     1. Update specific sections — tell me which sections need changes
     2. Full review — walk through the entire PRD for updates
     3. Cancel — no changes needed

3. Verify: Mode is set to create, update-section, full-review, or cancelled. If cancelled → stop.

Phase 1: Template & Reference Analysis

Step 1.1: Study Reference Materials

1. Check: Has the PRD template been analyzed in this session?
2. Skip: If yes, proceed to Step 1.2.
3. Act: Read and analyze:
   • PRD template — understand all required sections and structure
   • PRD example — understand quality standards, specificity level, writing tone
4. Verify: Template structure and quality expectations are understood.

Step 1.2: Create Section Checklist

1. Act: Build a checklist of all PRD sections from the template:

   PRD CHECKLIST:
   ├── [ ] 1. Overview (name, version, date, owner, summary)
   ├── [ ] 2. Vision & Mission
   ├── [ ] 3. Problem Statement (current state, pain points)
   ├── [ ] 4. Goals & Success Metrics (KPIs with targets)
   ├── [ ] 5. Target Users (personas, journey)
   ├── [ ] 6. Solution Overview (core solution, features P0/P1/P2)
   ├── [ ] 7. User Stories & Acceptance Criteria (epics, stories, ACs)
   ├── [ ] 8. Technical Considerations (architecture, requirements, constraints)
   ├── [ ] 9. Design Requirements (UI/UX, visual)
   ├── [ ] 10. Timeline & Milestones (phases, dependencies)
   ├── [ ] 11. Risks & Mitigations (table)
   ├── [ ] 12. Launch & Go-to-Market (strategy, marketing, support)
   └── [ ] 13. Post-Launch (monitoring, iteration plan)
   

2. Verify: All template sections are tracked.

Phase 2: Information Gathering

Step 2.1: Request Existing Documentation

1. Act: Ask the developer for any existing project materials:

   Before I start asking questions, do you have any existing documentation I should review? Examples: market research, user interviews, technical specs, business plans, competitive analysis, previous PRDs, wireframes, design mockups.

2. Act: If documentation provided, analyze each document:

   • Map which checklist sections are covered
   • Mark covered items in checklist
   • Extract relevant information

3. Verify: Documentation analysis complete. Remaining gaps identified.

Step 2.2: Iterative Hypothesis-Driven Questioning

For each uncovered checklist section, gather information using this pattern:

1. Act: Ask one specific question at a time with a plausible hypothesis based on available context:

   Based on [context/docs], I assume [specific hypothesis]. Is this accurate?

   Example questions:

   • "Based on the documentation, the primary problem is [pain point hypothesis]. Is this accurate?"
   • "The primary user persona would be [role hypothesis] who needs [need hypothesis]. Correct?"
   • "The must-have MVP features would include [3-4 feature hypotheses]. Are these the right priorities?"

2. Act: Wait for developer response (confirmation, correction, elaboration).

3. Act: Update checklist — mark section as covered when sufficient information gathered.

4. Act: Move to next uncovered section.

5. Verify: All checklist sections have sufficient information. If not, continue questioning.

Rules:

• Ask ONE question at a time — never batch multiple questions
• Always propose a reasonable assumption — never ask open-ended "what is X?"
• Validate understanding by summarizing complex responses
• Do NOT proceed to Phase 3 until all sections have information

Phase 3: PRD Creation

Step 3.1: Write PRD

1. Check: If mode is update-section, read existing PRD and modify only the target section. Skip to Step 3.2.
2. Act: Create (or overwrite template) at adoption/product/PRD.md:
   • Follow PRD template structure exactly
   • Match quality standard from PRD example
   • Replace all placeholders with gathered information
   • Use specific numbers, dates, and measurable criteria
   • Structure user stories and acceptance criteria per example format
3. Verify: PRD file created with all sections populated. No [placeholder] text remains.

Step 3.2: Self-Review

1. Act: Verify against quality standards:
   • Success metrics are measurable and time-bound with exact targets
   • Technical requirements are implementable and clearly articulated
   • User stories include clear acceptance criteria
   • Risks have corresponding mitigation strategies
   • All template sections completed
   • Terminology is consistent throughout
   • Writing tone matches example standard
2. Act: Fix any quality issues found.
3. Verify: PRD passes self-review.

Phase 4: Review & Approval

Step 4.1: Present for Review

1. Act: Present the complete PRD to the developer:

   I've completed the PRD. Please review each section and provide feedback on:

   • Accuracy of information
   • Missing details or requirements
   • Clarity and specificity
   • Alignment with your vision

   Please prioritize feedback as critical (must fix) or enhancement (nice to have).

2. Act: Apply feedback — update PRD file directly.

3. Act: Present updated sections for re-review.

4. Act: Repeat until developer approves.

5. Verify: Developer has explicitly approved the PRD.

Output Format

PRD COMPLETE:
├── Product:  [product name]
├── File:     [path to PRD file]
├── Mode:     [Created | Updated (sections: X, Y) | Full Review]
├── Sections: [N/N completed]
└── Status:   [Approved | Pending review]

Composition Interface

When composed by /pair-process-bootstrap:

• Input: /pair-process-bootstrap detects missing or template PRD and invokes /pair-process-specify-prd.
• Output: Returns path to the PRD file and approval status.
• /pair-process-bootstrap proceeds to next phase only after PRD is approved.

When invoked independently:

• Interactive: full Phase 0-4 flow. Developer commits changes when satisfied.

HALT Conditions

• Template not reviewed (Phase 1) — must understand structure before gathering information
• Example not analyzed (Phase 1) — must understand quality expectations
• Information incomplete (Phase 2) — do not create PRD with gaps
• Developer rejects PRD (Phase 4) — iterate until approved, never force completion

Graceful Degradation

• If PRD template not found, use the 13-section structure defined in Phase 1 Step 1.2 checklist as a minimal template.
• If PRD example not found, proceed without quality reference but warn: "No PRD example found — quality benchmarking unavailable."
• If adoption directory doesn't exist, create adoption/product/ and warn: "Created adoption directory — this appears to be a new project."

Notes

• Target file is always adoption/product/PRD.md — the single source of truth for product requirements.
• This skill modifies one file: the PRD. Changes should be committed by the calling skill or developer.
• Hypothesis-driven questioning reduces developer effort: confirm/correct is faster than open-ended answers.
• One question at a time prevents cognitive overload and ensures thorough coverage.