The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external, untrusted content (diffs or pull requests) which creates an attack surface for indirect prompt injection.\n- Ingestion points: Ingests code changes via the primary skill argument.\n- Boundary markers: No delimiters or safety instructions are present to differentiate the code being reviewed from the agent's instructions.\n- Capability inventory: The skill consists solely of instructions and does not invoke external tools or scripts.\n- Sanitization: No mechanism for validating or sanitizing the input data is provided.\n- [NO_CODE]: The skill contains only documentation and rules in markdown format, with no associated executable code.

code-review