The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and act upon data from external sources including steering files (steering/structure.md, steering/tech.md, steering/product.md) and various requirements documents in docs/requirements/. If these files contain malicious instructions, they could influence the agent's behavior, especially given its Bash, Write, and Edit capabilities.
Ingestion points: Reads project context from steering/ directory and requirements documents from docs/requirements/.
Boundary markers: No specific delimiters or safety warnings are established to distinguish between trusted instructions and potentially untrusted project data.
Capability inventory: The skill is granted Bash, Write, Edit, Read, Glob, and Grep tools, providing a significant execution and modification surface.
Sanitization: No sanitization or validation of external content is specified before interpolation into the agent's context.

performance-optimizer