What problem does it solve? Modifying the ai-badger framework is error-prone because edits must happen in features/<stack>/skills/ source directories while scaffolded copies in .ai-badger/skills/ and skills/ are overwritten by sync, and skipping the sync, freshness, or version steps breaks pre-commit gates and tests. This Skill encodes the verified edit-sync-test-build cycle and the pitfalls that repeatedly break contributors. ## Core Features & Use Cases - Source-of-truth workflow: Enforces editing features/<stack>/skills/<name>/ then running sync_plugin_skills.py, pytest, index_build.py --check, and the scaffold freshness guard before committing. - Subsystem recipes: Provides verified procedures for adding MCP catalog servers, framework hooks (Claude/Copilot/Hermes), project-local invariants, and skill scope registration in SKILL_SCOPES. - Release mechanics: Documents the VERSION bump, changelog, version_sync, and release_guard ordering, including the two-commit packaging forced by the pre-commit hook chain. - Use Case: When adding a new common MCP server to the catalog, follow the reference recipe covering meta.json, server.md line budget, tools.json tag vocabulary, availability gating, and the self-scaffold refresh with AI_BADGER_MCP_AVAILABILITY=all. ## Quick Start Ask the assistant to modify an ai-badger framework source file and run the full sync, test, and freshness-gate cycle before committing.