agent-laboratory-workflow

Agent Laboratory Workflow

agent-laboratory-workflow helps a coding agent deploy, configure, validate, and launch bounded AI4Math Auto-Research runs through Agent Laboratory.

Entry Point

Start with SKILL.md. It defines the deployment gates, API-key handling, research-topic intake, smoke checks, and recovery flow for incomplete Agent Laboratory runs.

Package Layout

• agents/: agent-facing metadata.
• references/: deployment and operational guidance.
• scripts/: helper commands for bootstrap, run inspection, and Codex research workspace handoff.
• templates/: experiment YAML templates for interactive auto-research runs.

Validation

Run from the repository root:

python3 -m unittest discover -s tests -v