/prd-stress-test

description: "Run multi-reviewer PRD stress test for build-readiness"
argument-hint: "<path-to-prd.md>"
allowed-tools: Read, Glob, Grep, Task
disable-model-invocation: true

PRD Stress Test

Run a multi-agent PRD review to answer: **"Is this PRD ready to build?"** Three reviewers analyze different dimensions in parallel, cross-reference findings, and deliver a consolidated review report.

Usage

/agent-teams:prd-stress-test path/to/prd.md

---

Overview

This command spawns three specialist reviewers to stress-test a PRD from different angles:

• **market-fit-reviewer**: Does this solve a real problem for a real market?
• **feasibility-reviewer**: Are the requirements clear and buildable?
• **scope-reviewer**: Is this appropriately sized for V1?

After parallel review, reviewers cross-reference findings to catch conflicts (e.g., scope-reviewer wants to cut a feature that market-fit-reviewer considers critical). The lead compiles everything into a consolidated review report with per-dimension scores and an actionable revision checklist.

**What you get:**

• Three independent review dimensions scored 1-5
• Blocking issues (must fix before building)
• Suggestions (nice to fix)
• Conflicts between reviewers
• Overall verdict: READY TO BUILD / NEEDS REVISION / MAJOR REWORK
• Specific revision checklist

---

PRD to Review

Read the PRD at: $ARGUMENTS

---

Instructions

Pre-Flight Check

1. Verify Agent Teams is available in your Claude Code version. If teammates cannot be spawned, display:

   This command requires Claude Code's Agent Teams feature.
   Check https://docs.anthropic.com/en/docs/claude-code for setup instructions.

If not available, stop.

2. Read the PRD file from the path in the "PRD to Review" section above.

• If file not found, display error and stop:

     Error: PRD file not found at [path].
     Please provide a valid path to a PRD markdown file.

---

Phase 1: PRD Intake

1. Parse the PRD content and identify key sections (features, requirements, target users, etc.).

2. Check for existing product context:

• Read `.claude/product-context/product-info.md` if it exists
• Read `.claude/product-context/competitive-landscape.md` if it exists

3. Display briefing:

── PRD Stress Test ────────────────────────────────────────

PRD: [file path]
Title: [extracted title or first heading]

Assembling your review team:
  1. market-fit-reviewer   → Market fit and differentiation
  2. feasibility-reviewer  → Technical feasibility and requirements clarity
  3. scope-reviewer        → Scope appropriateness and MVP sizing

Phase 1: Parallel Review (3 reviewers working simultaneously)
Phase 2: Cross-Reference (reviewers check each other's findings)
Phase 3: Consolidated Report (verdict with revision checklist)

Starting review...
───────────────────────────────────────────────────────────

---

Phase 2: Parallel Review

Spawn 3 teammates simultaneously using Agent Teams:

**Teammate 1: market-fit-reviewer**

Prompt: "Review this PRD for market fit. Score 1-5.

PRD CONTENT:
[full PRD content]

[Include any product context found in Phase 1]

Your job: Evaluate target user clarity, problem validation, value proposition,
differentiation, and market context. Use your market-fit-reviewer expertise.

Score 1-5: (1) No problem validation or market context. (2) Weak problem statement, no evidence. (3) Problem defined but market fit unvalidated. (4) Clear problem with evidence, competitive awareness. (5) Strong validated problem, clear differentiation, compelling market opportunity."

**Teammate 2: feasibility-reviewer**

Prompt: "Review this PRD for technical feasibility and requirements clarity. Score 1-5.

PRD CONTENT:
[full PRD content]

Your job: Evaluate requirements clarity, acceptance criteria, technical
feasibility, edge cases, and integration points. Flag every ambiguity.
Use your feasibility-reviewer expertise.

Score 1-5: (1) Requirements are vague, no acceptance criteria. (2) Some requirements defined but major gaps. (3) Requirements present but ambiguous edge cases. (4) Clear requirements with acceptance criteria, minor gaps. (5) Precise requirements, complete acceptance criteria, all edge cases covered."

**Teammate 3: scope-reviewer**

Prompt: "Review this PRD for scope appropriateness. Score 1-5.

PRD CONTENT:
[full PRD content]

Your job: Assess total scope, classify every feature as MUST-HAVE / CUT FROM V1 /
DEFER TO V2, identify scope creep, and estimate effort reduction from cuts.
Use your scope-reviewer expertise. Apply the 3-Feature MVP Rule.

Score 1-5: (1) Massive scope, no prioritization. (2) Overloaded scope with weak cuts. (3) Reasonable scope but unclear priorities. (4) Focused scope with clear MUST/CUT/DEFER. (5) Tight MVP scope, every feature justified, clear phasing."

Wait for all three reviewers to complete their reviews.

---

Phase 3: Cross-Reference

Send each reviewer the other two reviewers' findings to flag conflicts.

**To market-fit-reviewer:**

"Here are your fellow reviewers' findings. Flag any conflicts with your review.

FEASIBILITY REVIEW:
[feasibility-reviewer output]

SCOPE REVIEW:
[scope-reviewer output]

Specifically check:
- Are features you consider critical for differentiation marked 'CUT' by scope-reviewer?
- Do feasibility concerns affect market-critical features?
Flag conflicts and explain your position."

**To feasibility-reviewer:**

"Here are your fellow reviewers' findings. Flag any conflicts with your review.

MARKET FIT REVIEW:
[market-fit-reviewer output]

SCOPE REVIEW:
[scope-reviewer output]

Specifically check:
- Do features market-fit-reviewer considers critical have clear requirements?
- Do scope cuts remove technically risky components (positive) or create gaps?
Flag conflicts and explain your position."

**To scope-reviewer:**

"Here are your fellow reviewers' findings. Flag any co