The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for processing untrusted external data (code and artifacts) and performing actions based on that data, which constitutes an indirect prompt injection surface.\n- Ingestion points: The skill processes "code, diffs, artifacts, plans, or systems" as described in SKILL.md.\n- Boundary markers: Instructions for delimiters or ignoring embedded content are absent.\n- Capability inventory: Capabilities for file system modification and command execution are implied by the "apply the fix" and "verify it" steps in the instructions.\n- Sanitization: Specific sanitization or validation logic for the external inputs is not defined in the methodology.

revolver