The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted data from multiple sources, creating a potential surface for indirect prompt injection attacks.
Ingestion points: The skill is explicitly designed to retrieve data from the internet using WebSearch and WebFetch tools, as well as reading arbitrary local files using Read, Grep, and Glob tools.
Boundary markers: The instructions do not define boundary markers or delimiters to help the agent distinguish between its own system instructions and content retrieved from external sources.
Capability inventory: The skill has access to the Bash tool, which could be exploited to execute shell commands if the agent follows malicious instructions embedded in the external content it researches.
Sanitization: There is no evidence of sanitization, validation, or filtering logic applied to the external data before the agent processes it to generate reports or quizzes.

find-unknowns