github:dakesan/la-bench2025 · first committed 2025-11-21 · source · badge
Iteratively refine LA-Bench experimental procedures through validation and regeneration cycles. This skill should be used when improving generated procedures by ID, validating procedure quality against LA-Bench criteria, and managing the refinement loop between procedure-checker and procedure-generator skills. Triggered by requests to refine, improve, or validate LA-Bench procedures in output JSON
Listed, not yet scored. This skill has no public model test yet. It is in the scoring queue; submitted skills go first, then the most-installed.
1 commit to this file on GitHub (newest first).