Does this skill provide detailed feedback or just a pass/fail status?

It provides both. While the final output is a clean summary (Approved/Concerns/Rejected), each stakeholder identifies specific high-priority issues and concerns that need to be addressed.

How does this fit into the /cat:work workflow?

It is designed to run after the implementation phase is complete but before the final user approval, serving as a final automated validation step to catch edge cases from multiple professional viewpoints.

Can I manually choose which reviewers participate?

Yes, you can use the '## Force Stakeholders' section in your task's PLAN.md to explicitly include specific roles like legal, sales, or performance, regardless of the automated detection.

How does the skill decide which stakeholders to include in a review?

The skill uses a selection algorithm that scans your task's PLAN.md, commit messages, and specific file extensions to determine relevance. For example, changing CSS files automatically triggers a UX review, while changing auth logic triggers a Security review.

Multi-Perspective Stakeholder Review

Name: Multi-Perspective Stakeholder Review
Author: cowwoc

bycowwoc

•

セキュリティとテスト

Orchestrates specialized AI subagents to perform comprehensive quality reviews across architecture, security, performance, and legal domains.

The Stakeholder Review skill acts as an automated quality gate that simulates a cross-functional team review. By analyzing your task context, commit history, and modified files, it intelligently selects relevant reviewers—such as Security Architects, UX Designers, or Legal Experts—to validate implementation quality. It identifies domain-specific concerns early in the development cycle, providing an aggregated status report that ensures high-quality code and compliance before the final user approval gate.

主な機能

01Customizable review logic via manual stakeholder overrides in PLAN.md

02Context-aware stakeholder selection based on task types and file change patterns

03Aggregated quality gate reporting with clear status indicators (Approved, Concerns, Rejected)

04Multi-perspective analysis covering architecture, security, performance, UX, and legal

05Automatic detection of specialized needs for UI, API, and algorithm-heavy code

0656 GitHub stars

ユースケース

01Ensuring UI/UX consistency and accessibility for new frontend feature implementations

02Validating architectural integrity and security compliance during major refactors

03Running automated performance and quality checks before merging production-critical code

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add cowwoc/cat stakeholder-review

For use in Claude.ai and ChatGPT

Download Skill