Run three-perspective adversarial reviews on code changes and plans.
npx skills add https://github.com/sr-ai-dev/sr-harness --skill tribunal-sr-ai-dev
Please help me install this Agent Skill. Skill: tribunal Source: https://github.com/sr-ai-dev/sr-harness/tree/main/skills/tribunal Command: npx skills add https://github.com/sr-ai-dev/sr-harness --skill tribunal-sr-ai-dev
The tribunal skill helps you catch hidden risks, missing value, and feasibility gaps before you commit code or changes.
Run /tribunal with the path to your PLAN.md to receive a three-panel verdict and concrete revision actions.
High-intent search queries and answers about installing and using this skill.
An adversarial code review evaluates a pull request across risk, value, and feasibility dimensions to surface blind spots. This Skill runs three distinct analysis panels in parallel to produce a structured Tribunal Verdict with a scored decision matrix.
Yes, you can assess local uncommitted changes by running an adversarial analysis on your current diff. The review extracts must-fix items and evaluates feasibility gaps to help you decide whether to proceed, revise, or reject the changes.
A risk-value-feasibility matrix is a structured scoring method that evaluates a development plan from three adversarial perspectives. It synthesizes findings into an APPROVE, REVISE, or REJECT verdict, extracting actionable required changes based on panel findings.
Yes, this adversarial PR review works with GitHub pull requests, local plan files, and local diffs. It applies parallel risk, value, and feasibility analyses to the specified target to generate a final verdict with concrete revision actions.
To get actionable required changes from a code review verdict, the analysis categorizes findings into must-fix, should-address, and consider items. When the verdict is REVISE or REJECT, it provides specific required actions based on the synthesized panel findings.
Use a multi-perspective adversarial review when risk, value, and feasibility must be rigorously assessed before proceeding with a code change or plan. It surfaces hidden gaps and blind spots that a standard review might miss by running three independent analysis panels.
CircleCI-PublicReview GitHub PR or local diffs using team review prompt standards.
majiayu000Route user requests to registered skill modules and execute them in context.
Obsidian-OwlAudit automated test quality for assertion strength, boundary coverage, mock discipline, and error paths.
umairinayatLint, format, and statically analyze Node.js and Python code changes.
