The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill inventories and reads arbitrary files from a repository to perform its audit, exposing the agent to instructions potentially embedded in the repository content.
Ingestion points: Processes repository source, structure, manifests, and dependencies using tools like rg and direct file reads as described in the Process section.
Boundary markers: The instructions do not define specific delimiters or ignore instructions for the content being audited.
Capability inventory: The skill is strictly read-only and explicitly forbids editing, deleting, or committing.
Sanitization: No specific sanitization or validation of the ingested repository content is mentioned.

lean-audit