The Agent Skills Directory

[SAFE]: No malicious patterns detected. The skill's behavior is consistent with its stated purpose of enhancing other skills through automated review and user-approved modifications.
[PROMPT_INJECTION]: The skill processes external content from other skill files, creating an indirect prompt injection surface where instructions in a reviewed skill could attempt to influence the agent's behavior.
Ingestion points: Reads SKILL.md and supporting files from a user-provided directory.
Boundary markers: Absent; the skill does not use specific delimiters to isolate untrusted content from its own instructions.
Capability inventory: Has the ability to rewrite files on the local filesystem.
Sanitization: No explicit sanitization or validation of the ingested content is performed.
Mitigation: The risk is effectively mitigated by the mandatory user approval step required in the workflow before any changes are applied to the filesystem.

enhance-skill