2 added, 0 removed. Audit A to A.
# `.agents/skills/`
Vendored Agent Skills from
[muratcankoylan/Agent-Skills-for-Context-Engineering](https://github.com/muratcankoylan/Agent-Skills-for-Context-Engineering),
pinned by [`../context-engineering.lock.json`](../context-engineering.lock.json).
## Safety and maintenance
- Load `SKILL.md` progressively when its trigger matches; do not preload every
body into the context window.
- Treat `scripts/` and shell/Python snippets as examples. Never auto-execute
them; several accept caller-controlled paths or demonstrate credentialed
shell commands.
- Preserve relative paths when updating a skill.
+ - Do not place a skill from another upstream in this tree or absorb it by
+ changing only the recorded counts/digest; that would misstate provenance.
- Refresh through the audited installer at a pinned commit, reapply and record
local overlays, then update both upstream and vendored hashes.
- Run the upstream validators plus this repository's skill, runtime-sync, MCP,
Ruff, mypy, and test gates before changing the pin.