# Prd Review Panel > Run a 7-perspective multi-agent PRD review (Engineering, Design, Executive, Legal, UXR, Skeptic, Customer Voice) that catches gaps and surfaces conflicts before stakeholder review. Use when the user wants a PRD reviewed, asks for panel or multi-agent feedback, or runs /prd-review-panel. - Skill: `amit-t/prd-review-panel` (Agent Skill, multi-file: 4 files) - Install (CLI): `npx skillmds@latest add amit-t/prd-review-panel` - Raw SKILL.md: https://api.skillmd.com/api/skills/amit-t/prd-review-panel/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Product & Planning - Author: amit-t (https://skillmd.com/u/amit-t) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/amit-t/prd-review-panel --- ## Purpose Runs the 7 reviewers in parallel and explicitly challenges assumptions, not just gaps and conflicts. ## Usage - `/prd-review-panel` - Review a PRD with all 7 sub-agents - `/prd-review-panel [prd-name]` - Review specific PRD - `/prd-review-panel --perspectives "eng,design,exec"` - Review with subset of agents --- ## Context Routing **Check these files first:** 1. `outputs/prds/` - Active PRDs to review 2. `context-library/prds/` - Reference PRDs and past reviews 3. `sub-agents/` - The 7 reviewer personas 4. `context-library/strategy/` - Strategic context for executive review 5. `context-library/research/` - User research for UXR validation **Sub-agents available:** 1. **engineer-reviewer.md** - Technical feasibility, complexity, dependencies 2. **designer-reviewer.md** - UX/UI feedback, user experience 3. **executive-reviewer.md** - Strategic alignment, business impact 4. **legal-advisor.md** - Compliance, risk, regulatory concerns 5. **uxr-analyst.md** - User research synthesis, validation 6. **skeptic.md** - Devil's advocate, challenge assumptions 7. **customer-voice.md** - Simulate user perspective --- ## Workflow ### Step 1: PRD Selection 1. **If user specified PRD name:** - Look for it in `outputs/prds/` and `context-library/prds/` - If found: Proceed - If not found: List available PRDs, ask user to choose 2. **If no PRD specified:** - Scan `outputs/prds/` for recent PRDs (modified in last 30 days) - List them with: - File name - Title (from content) - Last modified date - Current stage (if indicated) - Prompt: "Which PRD do you want to review?" 3. **Read the full PRD:** - Load complete content - Note the current stage (Team Kickoff / Planning Review / XFN Kickoff / Solution Review / Launch Readiness) - Identify sections present (some PRDs may be incomplete) --- ### Step 2: Review Preparation **Extract key elements from PRD:** - **Problem statement** - What user pain are we solving? - **Hypothesis** - If we build X, then Y will happen because Z - **Strategic fit** - Why this vs other things? - **Non-goals** - What's explicitly out of scope? - **Success metrics** - How we measure success - **Rollout plan** - A/B test or full launch? - **Technical approach** - How we'll build it (if specified) - **UX design** - Mockups, flows, behavior examples - **Stakeholders** - Who needs to approve/support - **Risks** - Known concerns or open questions **Determine review focus based on stage:** - **Team Kickoff stage:** Focus on problem definition, strategic fit - **Planning Review stage:** Focus on scope, estimates, prioritization - **XFN Kickoff stage:** Focus on cross-functional alignment, dependencies - **Solution Review stage:** Focus on technical approach, UX design, edge cases - **Launch Readiness stage:** Focus on rollout plan, metrics, compliance --- ### Step 3: Spawn 7 Sub-Agents in Parallel **CRITICAL: Use single message with multiple Task tool calls for parallel execution.** For each sub-agent, create a Task using its prompt from [REVIEWER-PROMPTS.md](REVIEWER-PROMPTS.md), with the full PRD text substituted in. --- ### Step 4: Collect & Synthesize Reviews Once all 7 agents complete (wait for all Task outputs): **Read each review and extract:** 1. ✅ Strengths (what's working well) 2. ⚠️ Concerns or gaps (important but not blocking) 3. ❌ Blockers (must fix before proceeding) 4. 💡 Suggestions (improvements to consider) **Identify patterns:** - **Convergent feedback:** Multiple agents flagging same issue (high priority) - **Conflicting perspectives:** Agents disagree (requires PM judgment) - **Blind spots:** Issue only one agent caught (could be critical) **Categorize all feedback:** **Critical Blockers** (Must fix before next stage): - [Issue from Agent X] - [Issue from Agent Y] - **Why critical:** [Impact if not addressed] **Important Gaps** (Address before launch): - [Gap from Agent X] - [Gap from Agent Y] - **Why important:** [Risk or limitation] **Enhancements** (Consider for v1 or v2): - [Suggestion from Agent X] - [Suggestion from Agent Y] - **Value if added:** [Benefit] **Conflicting Perspectives** (Requires decision): - **Agent X says:** [Position] - **Agent Y says:** [Opposite position] - **PM decision needed:** [What to prioritize] --- ### Step 5: Generate Review Synthesis Create file: `outputs/prds/[prd-name]-review-synthesis.md` using the template in [SYNTHESIS-TEMPLATE.md](SYNTHESIS-TEMPLATE.md), with every placeholder filled in. --- ### Step 6: Output & Next Actions 1. **Save review synthesis file** 2. **Display summary:** ``` 7-agent review complete for [PRD]! ✅ Strong areas: [X] ⚠️ Critical blockers: [Y] 💡 Key recommendations: [Z] Overall: [Ready to proceed / Needs work] ``` 3. **Offer next steps:** - If blockers exist: "Want me to help update the PRD to address blockers?" - If conflicts exist: "Run `/decision-doc` to document the [Conflict] decision?" - If ready: "PRD looks solid! Ready for stakeholder review." - Always: "Review synthesis saved to [file path]" --- ## Tips for Best Results **When to run:** - After completing first draft (Team Kickoff stage) - Before stakeholder review (catch gaps privately first) - After major changes (validate new approach) - Before committing to build (final gate) **How to use the output:** - Fix blockers before proceeding - Discuss conflicts in next stakeholder meeting - Park enhancements for v2 (don't let perfect kill good) - Use as peer review before formal review **Common patterns:** - Engineering often conflicts with Design (speed vs polish) - Executive often conflicts with Customer (business vs user needs) - Legal often blocks what users want (compliance vs freedom) - **Your job as PM:** Make the call, document the trade-off --- ## Related Skills **Before this:** - `/prd-draft` - Create the PRD - `/user-interview` / `/user-research-synthesis` - Validate with research - `/impact-sizing` - Quantify value **After this:** - `/decision-doc` - Document key decisions - `/prd-draft` - Update PRD based on feedback - `/prototype` - Build based on feedback - `/launch-checklist` - Prepare for launch **Complements:** - `/competitor-analysis` - Inform strategic review - `/stakeholder-update` - Share review results **Iterative use:** Run review at each PRD stage (Team Kickoff, Planning, Solution, Launch) — each pass focuses on stage-appropriate concerns. --- ## Output Quality Self-Check Before presenting output to the PM, verify: - [ ] **All requested reviewer perspectives included:** The synthesis contains feedback from all 7 sub-agents (or all specifically requested agents), each with their own dedicated section - [ ] **Each reviewer has specific, actionable feedback:** Every reviewer section contains at least one concrete, actionable item (not generic praise like "looks good" or vague concerns like "needs more detail") - [ ] **Conflicting perspectives between reviewers explicitly flagged:** Any disagreements between agents (e.g., Engineering wants simplicity while Design wants richness) are called out in the "Conflicting Perspectives" section with both positions stated - [ ] **Synthesis section prioritizes feedback items:** The TL;DR and summary sections rank issues by severity (Critical Blockers > Important Gaps > Enhancements) with a clear recommended next step - [ ] **Feedback references specific PRD sections:** Each piece of feedback points to the exact section, requirement, or design element it applies to (e.g., "the rollout plan in Section 5" not "the rollout approach")