tribunal

Orchestrates external AI CLI audits with host arbiter synthesis.

Updated May 21, 2026

Search Tags:#security #audit #multi-model #code-review #parallel-execution #pre-production

Ready to run

npx skills add https://github.com/random1st/groundwork --skill tribunal-random1st

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: tribunal
Source: https://github.com/random1st/groundwork/tree/main/plugins/tribunal
Command: npx skills add https://github.com/random1st/groundwork --skill tribunal-random1st

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Tribunal enables independent audits of code by coordinating reviews from external AI CLIs while the host runtime acts as arbiter, ensuring unbiased verification of changes.

Core Features & Use Cases

Orchestrates multi-model reviews by selecting external providers (Claude, Codex, Gemini, Grok) and synthesizing results without the current runtime influencing auditors.
Supports Standard mode (2 auditors) and --critical mode (3 auditors) for higher assurance on security-sensitive code, pre-production deploys, and critical bug fixes.
Ideal for pre-merge reviews, security-sensitive changes, and cross-provider validation with auditable traceability.

Quick Start

Run /tribunal on a source file to trigger a standard two-auditor independent audit.

Frequently Asked Questions about tribunal

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I run a multi-model code review for security-sensitive changes?▼

Multi-model code review for security-sensitive changes is orchestrated by routing source files to external AI CLIs, using two independent auditors by default or three in critical mode, with the host runtime synthesizing the final verdicts.

Can I use external AI providers like Claude and Codex for independent code audits?▼

Yes, independent code audits can be performed by selecting external AI providers like Claude, Codex, Gemini, and Grok, ensuring the host runtime acts only as an arbiter to synthesize results without influencing the individual auditor reviews.

What is the difference between standard mode and critical mode in parallel code audits?▼

Standard mode in parallel code audits uses two external auditors to review changes, while critical mode engages three auditors to provide higher assurance for pre-production deploys and critical bug fixes before the arbiter synthesizes the verdicts.

How does an arbiter synthesize verdicts from autonomous AI auditors?▼

An arbiter synthesizes verdicts by collecting structured output formats from autonomously selected external AI auditors running in parallel, then merging these independent reviews into a final unbiased verification verdict.

When do I need a cross-provider validation review before a production deploy?▼

Cross-provider validation review is needed for pre-merge checks, pre-production fixes, and critical bug fixes where unbiased verification across multiple AI runtimes ensures higher assurance and auditable traceability before deployment.