mirror of
https://github.com/inference-sh/skills.git
synced 2026-09-14 18:43:24 +08:00
04f065fb4f
Agent self-regulation skill: read when stuck, looping, or on a mistake streak, then re-ground on verified observations and take one small step. Not topic-specific. Also adds an "Agent Practice" grouping to skills.sh.json and a README row.