The Agent Skills Directory

[SAFE]: The skill is purely instructional and focused on logical reasoning and documentation consistency. It does not perform network requests, execute external code, or access sensitive system files. \n- [SAFE]: Security-conscious design is evident in the inclusion of a shared 'GUARDRAILS.md' which mandates decision ownership by the user and prohibits any mutating actions without explicit approval. \n- [SAFE]: The skill explicitly defines its boundaries, handing off to other specialized tools for tasks like debugging or triage, which prevents scope creep and unexpected tool usage.

duck-grill