The Agent Skills Directory

[PROMPT_INJECTION]: The skill analyzes external SKILL.md files, creating a surface for indirect prompt injection where malicious content in those files could attempt to divert the agent's task.\n- [PROMPT_INJECTION]: Ingestion points: The skill processes frontmatter and content from all SKILL.md files found in the target directory.\n- [PROMPT_INJECTION]: Boundary markers: The instructions do not define strict delimiters or 'ignore' instructions for the parsed content to distinguish it from the skill's own logic.\n- [PROMPT_INJECTION]: Capability inventory: The agent is instructed to report findings and can apply file 'auto-fixes' if requested, providing a mechanism for malicious input to cause unintended file system changes.\n- [PROMPT_INJECTION]: Sanitization: No explicit content sanitization or validation of the input content is mentioned before it is processed by the agent.

enhance-skills