The Agent Skills Directory

[DATA_EXPOSURE]: The skill is configured to write product documentation to specific local paths: .dev/prd_doc.md and .dev/test_plan.md. These operations are limited to project documentation within a hidden directory commonly used for development metadata and do not involve access to sensitive system files or credentials.
[PROMPT_INJECTION]: The skill contains robust internal constraints that instruct the agent to ignore technical implementation details (code, APIs, architecture) and stick strictly to a product-facing PRD format. These instructions serve as a protective boundary against attempts to repurpose the agent for unauthorized code generation.
[INDIRECT_PROMPT_INJECTION]:
Ingestion points: The skill processes external feature requests and requirements documents provided as user input.
Boundary markers: No explicit XML/delimiter boundaries are defined for the input data.
Capability inventory: The skill has the capability to write Markdown files to the local file system (.dev/prd_doc.md, .dev/test_plan.md).
Sanitization: The skill relies on strict output format enforcement (Markdown sections) rather than explicit sanitization of the input content. The risk is minimized by the non-executable nature of the output format.

prd-analysis