yogirk

yogirk / agent-council

Public

A skill that convenes a panel of CLI-based AI agents (Claude Code, Codex, Gemini CLI) to deliberate on engineering problems through structured multi-stage discussion, peer review, and synthesis

15
1
100% credibility
Found Mar 31, 2026 at 15 stars -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Agent Council assembles multiple AI agents to independently deliberate on questions, review each other, and produce a synthesized recommendation as a group verdict.

How It Works

1
📰 Discover Agent Council

You hear about a helpful tool that gathers a team of smart AI advisors to discuss tough decisions together.

2
⚙️ Get it ready

Follow easy steps to set up the council using tools you already have for chatting with AIs.

3
📝 Prepare your question

Write down a big choice you're facing, like picking the best way to store data in your project.

4
🚀 Ask the council

Share your question and watch as different AI members think independently and share their views.

5
💬 Follow the discussion

See the agents review each other's ideas and the leader pulls it all together into a clear summary.

6
📊 Explore the verdict

Open a simple webpage to dive into agreements, differences, and the final recommendation.

✅ Decide with confidence

Use the team's wise advice to make better choices for your project, and revisit anytime things change.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 15 to 15 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 agent-council?

Agent Council AI is a TypeScript tool that assembles a panel of CLI-based LLM agents—Claude Code, OpenAI Codex CLI, and Gemini CLI—to debate engineering decisions like database choices or job queues. Invoke it via `/council "Postgres or DynamoDB?"` in any of those CLIs, and it runs parallel independent opinions, optional peer reviews, then a chairman synthesis highlighting consensus, divergence, and confidence. Sessions save to disk with an interactive HTML viewer for replay, revisit with fresh context, or outcome logging.

Why is it gaining traction?

Unlike API-only agent councils, this leverages your existing CLI agent subscriptions for zero extra cost, letting agents grep codebases and inspect migrations for grounded advice. Benchmarks show it surfaces 2x more considerations than solo agents, with proactive nudges in CLIs and "living decisions" via `/council-revisit`. Developer-friendly slash commands and macOS/Linux setup via `npx cliagent-council` make agent council github experiments addictive.

Who should use this?

Backend engineers debating event sourcing or auth flows, tech leads comparing queues like BullMQ vs Agenda, or solo devs needing multi-model checks on scaling trade-offs. Ideal for Node.js/TS teams with Claude Code or Codex CLI already running, especially architecture reviews or debugging latency spikes.

Verdict

Try it if you use CLI agents—solid docs, evals, and viewer punch above 15 stars and 1.0% credibility. Early maturity means watch for Windows/roadmap gaps, but MIT license lowers risk for agent council llm tinkering.

(187 words)

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.