LLM Council — Claude Code skill: what it does, how to install
LLM Council coordinates multiple LLMs across three stages: independent responses, anonymized peer grading, and chairman synthesis.
no license 5 installs ★ 38 updated 2025-12-05
npx skills add shuntacurosu/llm_council_skill --skill llm-council -a claude-codeWhat LLM Council does
The skill queries several AI models at once, strips identifiers from their submissions, and lets them rank each other's outputs. A designated chairman model then gathers the strongest points into a unified summary. During code updates, the workflow can isolate each model's changes via Git worktrees to evaluate competing diffs safely.
Execution requires Python 3.8 or newer along with python-dotenv and loguru. Model endpoints and credentials reside in scripts/.env, and OpenCode CLI must be installed. Built-in flags allow running a terminal dashboard via --dashboard, reviewing diffs through --dry-run, or applying the top-rated patch automatically using --auto-merge.
What it helps with
- Gather contrasting perspectives on architecture decisions from multiple LLMs
- Run blind peer reviews on code patches across different models
- Evaluate competing refactoring proposals in isolated Git worktrees
Example prompts for the agent
- Ask the council to evaluate caching methods:
python scripts/run.py council_skill.py "What's the optimal caching strategy?" - Inspect competing bug fixes without applying changes:
python scripts/run.py council_skill.py --dry-run "Fix the bug in buggy.py" - Track the council debate live in the terminal:
python scripts/run.py cli.py --dashboard "Evaluate database migration plans"
How to install LLM Council
1Skills CLI (Vercel)
npx skills add shuntacurosu/llm_council_skill --skill llm-council -a claude-code2Manually: put the files in place
~/.claude/skills/shuntacurosu-llm-council/SKILL.mdCopy the whole skill folder; for one project use .claude/skills/shuntacurosu-llm-council/
1Skills CLI (Vercel)
npx skills add shuntacurosu/llm_council_skill --skill llm-council -a codex2Manually: put the files in place
.agents/skills/shuntacurosu-llm-council/SKILL.mdFor all projects use ~/.agents/skills/
1Skills CLI (Vercel)
npx skills add shuntacurosu/llm_council_skill --skill llm-council -a cursor2Manually: put the files in place
.cursor/skills/shuntacurosu-llm-council/SKILL.mdCursor also reads .agents/skills/ and .claude/skills/
1GitHub CLI
gh skill install shuntacurosu/llm_council_skill shuntacurosu-llm-council2Manually: put the files in place
.github/skills/shuntacurosu-llm-council/SKILL.md1Skills CLI (Vercel)
npx skills add shuntacurosu/llm_council_skill --skill llm-council -a gemini-cli2Platform CLI
gemini skills install https://github.com/shuntacurosu/llm_council_skillManually: put the files in place
.devin/skills/shuntacurosu-llm-council/SKILL.mdThe legacy .windsurf/skills/ path still works
Skills CLI (Vercel)
npx skills add shuntacurosu/llm_council_skill --skill llm-council -a opencodeClaude.ai
Zip the folder and upload it in Settings → Capabilities → Skills
Step by step, with commands for each platform: How to install skills in Claude Code: five ways
Questions about LLM Council
How to install LLM Council in Claude Code?
Run npx skills add shuntacurosu/llm_council_skill --skill shuntacurosu-llm-council -a claude-code. The definition will be placed in ~/.claude/skills/shuntacurosu-llm-council/SKILL.md.
What does LLM Council do?
It runs a three-stage deliberative process where multiple LLMs answer a prompt, anonymously grade peer outputs, and combine their findings under a chairman model.
What prerequisites are needed to run LLM Council?
You need Python 3.8 or higher, loguru, python-dotenv, and a configured OpenCode CLI. Specify your chosen models inside scripts/.env and inspect setup details with python scripts/run.py council_skill.py --setup.