The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes arbitrary external text ('output to evaluate') which serves as a potential vector for indirect prompt injection.
Ingestion points: Untrusted data is ingested via the skill's main argument as the content to be evaluated.
Boundary markers: The skill instructions do not specify explicit delimiters or 'ignore' instructions for the content being analyzed.
Capability inventory: The skill is configured to use the Read and Grep tools, limiting the potential impact of an injection to information retrieval within the allowed scope.
Sanitization: No specific sanitization or filtering of the input content is mentioned in the skill definition.

evals