Skip to content

Latest commit

 

History

History
102 lines (70 loc) · 3.04 KB

File metadata and controls

102 lines (70 loc) · 3.04 KB

Installation

OpenClaw (Recommended)

From ClawHub

openclaw skills add quorum

That's it. Quorum is now available to your agent. Ask it to run a quorum check on any file.

From Source

Clone the repo and point OpenClaw at it:

git clone https://github.com/SharedIntellect/quorum.git
cd quorum

The reference implementation lives in reference-implementation/. You can run it directly:

cd reference-implementation
pip install -r requirements.txt
python -m quorum.cli run examples/sample-config.yaml --rubric examples/rubrics/research-quality.json

Verify It Works

The quickest way to confirm everything is set up:

# From the reference-implementation directory:
python -m quorum.cli run examples/sample-config.yaml --rubric examples/rubrics/research-quality.json

You should see:

  1. First-run config setup (if no quorum-config.yaml exists) — it'll ask your preferred models
  2. Critics spawning and evaluating the sample artifact
  3. A structured verdict with findings and evidence citations

If you see a PASS, PASS_WITH_NOTES, or FAIL verdict with cited evidence, Quorum is working.

Configuration

After first run, Quorum creates quorum-config.yaml in your working directory. Edit it to change models, depth profiles, or critic selection:

models:
  tier_1: anthropic/claude-sonnet-4-6
  tier_2: anthropic/claude-sonnet-4-6

depth: standard    # quick | standard | thorough

See CONFIG_REFERENCE.md for full options and MODEL_REQUIREMENTS.md for supported models.

Requirements

  • Python 3.10+
  • An API key for at least one supported model provider (Anthropic, OpenAI, or Google)
  • OpenClaw (optional but recommended — handles agent orchestration automatically)
  • LiteLLM (installed via requirements.txt — provides universal model access)

Troubleshooting

"No model configured" — Run quorum run once and follow the first-run setup prompts, or manually create quorum-config.yaml (see CONFIG_REFERENCE.md).

"API key not found" — Set your provider's API key as an environment variable:

export ANTHROPIC_API_KEY=sk-ant-...
# or
export OPENAI_API_KEY=sk-...

Critics timing out — Large artifacts with thorough depth can take 45-90 minutes. Try quick depth first to verify setup, then scale up.

"Model not supported" — Check MODEL_REQUIREMENTS.md for compatible models. Quorum requires models with strong reasoning and tool-use capabilities.

Updating

ClawHub

openclaw skills update quorum

From Source

cd quorum
git pull origin main
pip install -r reference-implementation/requirements.txt

⚖️ LICENSE — Not part of the operational specification above. This file is part of Quorum. Copyright 2026 SharedIntellect. MIT License. See LICENSE for full terms.