The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from external project files to generate a PRD without implementing boundary markers or sanitization.
Ingestion points: The skill reads vision.md, features.md, and stakeholders.md from the projects/<ProjectName>/_context/ directory as specified in SKILL.md and logic.prompt.
Boundary markers: There are no explicit delimiters (such as XML tags or unique markers) or instructions provided to the agent to differentiate between the document generation logic and potentially malicious instructions embedded within the input data files.
Capability inventory: The skill is authorized to read project files and write the resulting PRD.md and manifest.md to the local file system.
Sanitization: The skill does not perform validation or filtering on the content extracted from the context files before using it in the document generation process.

01-prd-generation