The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied specification files, which creates a surface for instructions embedded in those files to influence the behavior of the review agents.
Ingestion points: The skill identifies and instructs sub-agents to read feature specification files (e.g., plan.md, requirements.md, design.md, tasks.md) located in the docx/features/ directory.
Boundary markers: The skill does not define explicit delimiters or instructions to the sub-agents to ignore potentially malicious commands embedded within the specification content.
Capability inventory: The skill uses the Read, Bash, Glob, and Agent tools to locate, read, and audit the files.
Sanitization: There is no evidence of content validation or sanitization of the specification files before they are read by the agents.

spec-review