Can I use Tribunal for GitHub Pull Requests?

Yes, by using the --pr flag followed by the PR number, Tribunal can automatically fetch the diff and description from GitHub to perform its review.

What happens if the agents disagree on a verdict?

The skill includes a 'Contention Points' section that specifically highlights where the agents have conflicting views (e.g., high value vs. high risk), providing the context you need to make the final call.

What is the difference between the Risk and Feasibility agents?

The Risk analyst (Codex-powered) looks for what could go wrong or break, while the Feasibility checker evaluates if the solution is practically buildable within your specific codebase patterns and dependencies.

How does the Tribunal skill work?

It launches three specialized sub-agents simultaneously—one focusing on risks, one on business value, and one on technical feasibility—then merges their findings into a single comprehensive report with a final verdict.

Tribunal 3-Perspective Review

Name: Tribunal 3-Perspective Review
Author: team-attention

byteam-attention

•

119

•

Collaboration Tools

Conducts a multi-perspective adversarial review of code, plans, and pull requests to assess risk, value, and feasibility.

Tribunal is a specialized Claude Code skill designed for rigorous decision-making and quality assurance through adversarial review. It orchestrates three distinct AI agents—a risk analyst, a value assessor, and a feasibility checker—to perform a parallel evaluation of your work. By synthesizing these diverse viewpoints into a unified verdict (Approve, Revise, or Reject), it helps developers identify hidden technical debt, validate project requirements, and ensure that every change adds genuine value while remaining practical to implement within the existing codebase.

Key Features

01Direct integration with local files, Git diffs, and GitHub Pull Requests

02Unified verdict synthesis with clear actionable requirements

03Parallel 3-agent adversarial review architecture

04Automated scoring for Risk, Business Value, and Technical Feasibility

05Conflict resolution that identifies contention points between agents

06119 GitHub stars

Use Cases

01Pre-deployment risk assessment for critical architectural changes

02Conducting deep-dive code reviews on complex or high-stakes Pull Requests

03Validating project plans against requirements and technical constraints

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add team-attention/hoyeon tribunal

For use in Claude.ai and ChatGPT

Download Skill