Security audit
Security checks for vulnerabilities and agentic risk
This is a markdown-only orchestration skill that routes work to specialist agents and includes user checkpoints, with no evidence of hidden execution, persistence, credential access, or data exfiltration.
Install this if you want an agent-planning and routing helper. Review its proposed specialist routing and checkpoints carefully for sensitive work such as security, deployments, authentication, or large refactors, because the triggers are broad.
The skill description defines the orchestrator in very broad terms and encourages use across many domains without clear limits on what it should handle directly versus when it should defer, ask clarifying questions, or refuse. That ambiguity increases the chance of overreach, inappropriate routing, and execution of risky workflows based on incomplete user intent.
The routing triggers are broad, keyword-based, and overlapping, which can misclassify ordinary user requests and invoke overly powerful or irrelevant specialists. In an orchestrator skill, this can cause unintended delegation, scope expansion, and unsafe handling of sensitive tasks such as security, architecture, or code changes without sufficiently precise intent determination.
No suspicious patterns detected.
