The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes project-specific external data such as ticket.md, source code, and documentation to generate specifications.
Ingestion points: The agent is instructed to read ticket.md, the brief, AGENTS.md, CLAUDE.md, and source code files.
Boundary markers: No specific boundary markers or safety delimiters are defined for the external data ingestion.
Capability inventory: The skill performs file system read/write operations and triggers other workflow skills like flow-review and flow-plan.
Sanitization: The instructions do not include data validation or sanitization for the content read from external files.

flow-spec