Security audit
Security checks for vulnerabilities and agentic risk
This is a minimal instruction-only skill with broad positioning but no artifact-backed code execution, file access, network use, persistence, or hidden behavior.
Install only if you want a very general planning and orchestration prompt. Because it is broad and sparse, keep normal agent permissions and approvals in place for sensitive work such as legal, medical, financial, account, or system-changing tasks.
The description presents the skill as a general operating-system layer that routes goals to skills and verifies results, but it does not define clear activation boundaries, trust constraints, or disallowed contexts. In an agent ecosystem, this kind of broad, always-applicable framing can cause over-selection, priority hijacking, or unsafe delegation into other skills without adequate guardrails.
The body text claims broad orchestration capabilities—routing goals, executing with checkpoints, and verifying results—without specifying decision rules, exclusions, or operator controls. Because this is an instruction-only meta-skill intended to sit above other skills, ambiguous scope is more dangerous here than in a narrowly scoped utility skill: it can influence many downstream actions and create unsafe or conflicting behavior across the whole agent stack.
No suspicious patterns detected.
