How does the skill handle conflicting reviews from different models?

The skill synthesizes reports by deduplicating issues, highlighting high-confidence bugs found by multiple reviewers, and flagging any gate that fails for at least one model.

Which AI models does the Spec Review skill use?

It utilizes a combination of native Claude models (Opus and Sonnet) along with external models like GPT-5.2 and Gemini-3 Pro via the OpenCode tool for diverse perspectives.

Can I run this skill on existing specifications?

Yes, you can use the /spec.review command followed by a spec name, or let it automatically detect the most recent draft in your specs folder.

Multi-Agent Spec Reviewer

Name: Multi-Agent Spec Reviewer
Author: srnnkls

bysrnnkls

0•

협업 도구

Performs comprehensive, multi-perspective technical specification reviews using parallel AI subagents to ensure project feasibility.

The Spec Review skill provides a robust validation layer for technical documentation by dispatching parallel subagents—including Claude, GPT, and Gemini—to analyze project specifications from multiple viewpoints. By evaluating documents against four critical gates—Completeness, Consistency, Feasibility, and Clarity—it identifies potential bottlenecks, ambiguous requirements, and missing dependencies before implementation begins. It synthesizes diverse model perspectives into a unified report, facilitating automated documentation updates and resolving edge cases through interactive user clarification to significantly reduce downstream development risks.

주요 기능

01Integrated workflow that bridges the gap between spec creation and task implementation.

02Parallel multi-agent analysis using Claude, GPT, and Gemini models simultaneously.

03Four-gate validation system checking Completeness, Consistency, Feasibility, and Clarity.

04Automated issue synthesis and de-duplication across multiple AI reviewers.

050 GitHub stars

06Interactive clarifying questions with automated updates to validation.yaml files.

사용 사례

01Identifying edge cases and logic gaps in complex system architecture documentation.

02Validating a newly created technical specification before moving to the implementation phase.

03Ensuring cross-document consistency between tasks, context, and requirement files.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add srnnkls/tropos spec-review

For use in Claude.ai and ChatGPT

Download Skill