The Agent Skills Directory

[SAFE]: The skill utilizes limited tool permissions (Read, Glob, Grep) defined in the YAML frontmatter, which restricts the agent's ability to perform destructive actions, write files, or execute network requests.- [SAFE]: The available agents and expertise levels described (e.g., security-auditor, penetration-tester) are persona definitions for security analysis and testing within a development context.- [SAFE]: No obfuscation, hidden instructions, or persistence mechanisms were detected in the skill body or metadata.- [SAFE]: The orchestration patterns provided are instructional guidelines for using native platform capabilities and do not introduce external security risks.

parallel-agents