The Agent Skills Directory

[SAFE]: The skill's primary function is to analyze and generate documentation specifically intended for AI agents (e.g., AGENTS.md, llms.txt). All analyzed files are consistent with this stated purpose.
[COMMAND_EXECUTION]: The evals/run-static-checks.sh script executes shell commands and a Python-based static checker (check-skill-static.py). This is standard behavior for automated skill validation and is restricted to the development/evaluation environment.
[SAFE]: No instances of prompt injection, data exfiltration, or hardcoded credentials were found. The skill does not perform network operations beyond the standard capabilities of the agent platform.
[SAFE]: The logic for sub-agent dispatch defined in references/subagent-dispatch.md follows standard multi-agent orchestration patterns for complex tasks like documentation auditing.
[SAFE]: The skill ingests untrusted data from target repositories (like AGENTS.md), which is an inherent part of its auditing function. While this presents a surface for indirect prompt injection, the skill includes structured rubrics and playbooks that guide the agent to evaluate content neutrally.

agent-docs