The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external, untrusted content which may contain instructions designed to deceive the agent.
Ingestion points: Step 1 in SKILL.md identifies that the review target and requirements can be provided via path, URL, pasted text, or IDE context.
Boundary markers: The instructions lack explicit requirements for the agent to use delimiters or protective prompting to isolate untrusted content from its operational instructions.
Capability inventory: The skill includes a file-writing capability in Step 5 (Optional plan update), allowing it to modify implementation plans or code based on the preceding analysis.
Sanitization: The skill does not specify any sanitization, filtering, or validation steps for the content fetched from URLs or read from the filesystem before processing it.

peer-review