The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to read and cross-reference the project's source code and existing documentation to ensure consistency with the domain model. This represents an indirect prompt injection surface as these files could contain embedded instructions designed to influence the agent's behavior during the modeling process.
Ingestion points: Processes content from CONTEXT.md, CONTEXT-MAP.md, and various project source code files.
Boundary markers: No specific delimiters or safety warnings for handling embedded instructions within these files are defined.
Capability inventory: The skill is capable of modifying CONTEXT.md and creating new Architectural Decision Record (ADR) files in the docs/adr/ directory.
Sanitization: No explicit sanitization or validation of the ingested file content is performed.

domain-modeling