What problem does it solve? Managing agent skills across nested canonical directories, flat .agents/skills/ and .claude/skills/ symlink overlays, and skills-lock.json is error-prone; this Skill enforces the correct topology and safe npx skills workflows for adding, editing, renaming, promoting, and removing skills. ## Core Features & Use Cases - First-party skill lifecycle: Add, edit, rename, promote, demote, or remove repo-local skills while preserving the canonical skills/<disposition>/<family>/<identity>/ layout and symlink invariants. - GitHub-sourced vendored skills: Install and update third-party skills as real vendored directories without creating canonical directories or clobbering curated lockfile selections. - Safe validation: Avoids the destructive npx skills check command, relying instead on npx skills list --json, readlink checks, and bounded reference searches. - Use Case: When adding a new internal skill, follow the workflow to bootstrap with npx skills add, replace the bootstrap copy with a symlink to the canonical destination, normalize the lock entry, and verify both overlay links before staging. ## Quick Start Ask the agent to add a new first-party skill at an approved destination using npx skills and wire up the .agents and .claude symlinks.
