berrzebb

berrzebb / consensus-loop

Public

Claude Code plugin — multi-agent orchestration with cross-model audit gate. Orchestrator distributes tasks to isolated worktree workers, GPT/Codex audits, HITL retrospective enforces quality. Hooks + Skills + Agents in one plugin.

ai-agents claude-code claude-code-plugin code-review consensus
11
3
100% credibility
Found Mar 19, 2026 at 11 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
JavaScript
AI Summary

A plugin for an AI coding tool that enforces a structured process where one AI implements code, another audits it independently, and team reflections ensure continuous improvement.

How It Works

1
🔍 Discover the helper

You hear about consensus-loop, a smart buddy for your AI coding assistant that makes sure code gets double-checked before using it.

2
📦 Add it easily

You simply add this review tool to your AI assistant, and it sets up all the helpful parts automatically.

3
⚙️ Set your preferences

You pick a spot to track feedback and choose tags like 'review needed' or 'approved' to guide the process.

4
🚀 Start a coding adventure

You tell the orchestrator to plan and kick off the session, and it guides multiple AI helpers to build and check your code safely.

5
🔍 Watch the review magic

One AI creates the code while another independently audits it for bugs, tests, and quality, ensuring nothing slips through.

6
💡 Learn and tweak

You join a quick team chat to reflect on what worked well and fix any needed changes before moving forward.

✅ Code ships perfectly

Your improved, thoroughly vetted code gets added to the project, and you're ready for the next round with smarter results.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 11 to 11 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is consensus-loop?

Consensus-loop is a JavaScript Claude Code plugin that orchestrates multi-agent AI coding with a built-in cross-model audit gate—Claude implements in isolated git worktrees, while GPT or Codex reviews changes independently, blocking merges without approval tags like [APPROVED]. It solves AI tools' bug-prone self-review by enforcing peer audits and human-in-the-loop retrospectives via simple claude code install commands: `claude plugin install consensus-loop@berrzebb-plugins`. Users get CLI skills like `/consensus-loop:orchestrator` for task distribution and parallel workers, plus hooks for automatic quality checks on edits.

Why is it gaining traction?

It stands out by using structure over prompts—nothing ships without consensus from a different model, catching bugs AI misses in self-review, with features like RTM traceability and 8 verification categories (tests, coverage, security). Developers hook it into claude github integration or claude code skills for seamless claude github code review, running async audits and session gates without blocking workflows. The parallel worktree agents and deterministic tools deliver production results, like fixing cross-track issues in large JS codebases.

Who should use this?

AI-heavy JS teams building complex apps, like those managing 100+ file workflows with claude code cli or claude github plugin. Solo devs on claude code windows seeking claude code free quality gates for claude github copilot outputs. Orchestrators tired of manual merges in claude github actions or claude code review pipelines.

Verdict

Early alpha with 11 stars and 1.0% credibility score—docs shine with real-world screenshots, but lacks broad testing; try the lightweight audit gate first via claude code download. Solid for Claude users wanting consensus loop rigor, skip if not deep in claude github mcp or claude code pricing experiments.

(198 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.