The Agent Skills Directory

[SAFE]: Analysis of all skill files, including orchestrator and sub-agent instructions, revealed no malicious code or command injection vectors.
[SAFE]: The skill demonstrates excellent security hygiene by incorporating a dedicated security review agent and providing patterns for threat modeling and AI safety.
[SAFE]: Credential management is handled through secure best practices, using environment variables and placeholders for sensitive information.
[SAFE]: Resource access via allowed tools (Bash, Read, etc.) is strictly bounded to architectural analysis and project metadata discovery.

system-architecture