Plugin

colosseum

proven wrong by search results, and earn scores based on accuracy and independence — not agreement. A built-in Devil's Advocate challenges the final synthesis before it's delivered. How it works: 1. Builds a fact base from opposing-angle web searches 2. Each agent declares an independent position (blind draft — no agent sees others before submitting) 3. Agents debate up to 5 rounds: Prosecutor attacks, Defender responds, Adverse Witness probes shared blind spots 4. Role-switch gate fires when agents start citing the same sources or hit deadlock 5. Devil's Advocate challenges the final answer; a quantitative score (Accuracy · Evidence · Independence) determines whether it stands Works with Gemini CLI, Codex, aichat, or llm. Falls back to Claude's four internal roles (Pragmatist / Skeptic / Idealist / Realist) with zero setup.

• Published source
• Repository

Analysed

11 October 2026

Last observed

12 October 2026

What the code does

• Runs shell commands Code that starts system commands on the machine it runs on.
• Runs hooks automatically Hooks that execute without being called explicitly.

Having a capability is not wrongdoing; many tools need these to do their job. It tells you what to check before you install.

This is public metadata, not a safety certification or complete assessment. Static analysis cannot establish what a tool does at runtime.