What problem does it solve? Projects scaffolded by ai-badger fall behind the framework as new versions ship, and nobody knows what changed or whether local edits will be overwritten. This Skill reports exactly what drifted, backs up .ai-badger/, and re-scaffolds from the project's existing config.json without re-asking setup questions. ## Core Features & Use Cases - Drift detection and reporting: Compares the project's manifest against the framework's current content and reports changed, removed, orphaned, locally modified, and newly added catalog items as structured JSON. - Safe re-scaffolding: Backs up .ai-badger/ before rewriting managed files, preserves seed-once files (state.json, markers-context.json, model.json) and content inside keep-region markers, and re-embeds skill extensions automatically. - Hygiene reporting: Surfaces competing framework copies, unused delivered skills with usage evidence windows, and orphaned Hermes skill namespaces — pruning only on explicit user approval. - Use Case: A drift notice appears after ai-badger 0.60.0 ships. Run the refresh to see which skills changed upstream, review the git diff of regenerated CLAUDE.md and AGENTS.md files, and commit the updates. ## Quick Start Ask the agent to refresh ai-badger in this project and show what changed before committing.