The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided design documents and project files, which serves as a surface for indirect prompt injection where malicious instructions could be embedded in the content being reviewed.
Ingestion points: Target design documents, CLAUDE.md, and narrative files located in the design/ directory are read into the agent context.
Boundary markers: The instructions do not specify explicit delimiters or "ignore instructions" wrappers for the ingested document content.
Capability inventory: The skill is authorized to use Read, Write, Edit, and Task (sub-agent delegation) tools, allowing it to modify project files and spawn independent agent sessions.
Sanitization: The skill does not perform specific sanitization or filtering of the ingested document content before processing it or passing it to specialist sub-agents.

design-review