The Agent Skills Directory

[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection (Category 8) as it processes untrusted data from the local repository and user input to generate instructions for downstream agent tasks.
Ingestion points: The agent ingests free-form user intake text and repository-specific documentation files, including CONTEXT.md, CONTEXT-MAP.md, and Architectural Decision Records (ADRs) found in docs/adr/*.
Boundary markers: The skill instructions do not prescribe the use of boundary markers or clear delimiters to differentiate between processed data and instructional content in the generated output.
Capability inventory: The skill is designed to perform read-only codebase scans using grep and ripgrep (rg) and has the capability to write markdown files to the docs/prompt/ directory.
Sanitization: No sanitization or validation steps are defined to filter potentially malicious instructions embedded within the user requirements or the scanned documentation files before they are incorporated into the final generated prompt.

feature-prompt