The Agent Skills Directory

[SAFE]: The skill is purely instructional and focused on documentation. It does not perform network operations, access sensitive system paths, or execute shell commands.\n- [SAFE]: The instructions explicitly include security-conscious boundaries, advising the agent to 'Never' commit secrets and to 'Ask first' before adding new dependencies, which aligns with the principle of least privilege.\n- [SAFE]: The skill processes structured documentation input ('Confirmed Intent') to guide its output but does not interpolate this data into executable contexts or dangerous tool calls.

spec