The Agent Skills Directory

[PROMPT_INJECTION]: Indirect prompt injection surface identified due to external data ingestion.\n- Ingestion points: Data is brought into the agent context via WebFetch and WebSearch tools defined in SKILL.md.\n- Boundary markers: The skill lacks explicit delimiters or instructions to treat external data as untrusted, which could allow embedded prompts to be obeyed.\n- Capability inventory: The skill utilizes Read, Glob, Grep, WebFetch, and WebSearch. It does not have access to tools for writing files or executing shell commands.\n- Sanitization: No input validation or filtering is applied to the content fetched from the web.

prd