The Agent Skills Directory

[SAFE] (SAFE): No malicious patterns, credential exposure, or obfuscated payloads were detected in the skill metadata or body.
[NO_CODE] (INFO): The skill lacks executable scripts or binary files, significantly reducing the potential for runtime attacks.
[INDIRECT_PROMPT_INJECTION] (LOW): The skill instructs the agent to read from local reference files (e.g., references/patterns.md). Although this introduces a data ingestion surface, the skill has no high-privilege capabilities such as shell execution or network access to exploit.

llm-fine-tuning