The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes project-specific documentation files (CONTEXT.md, CONTEXT-MAP.md, and ADRs) and user input to manage the domain model.
Ingestion points: Reads existing CONTEXT.md, CONTEXT-MAP.md, and architectural decision records from the repository.
Boundary markers: The instructions do not define specific delimiters for untrusted content or warn the agent to ignore instructions embedded in the domain model files.
Capability inventory: The agent's capabilities are limited to reading and writing markdown-based documentation files.
Sanitization: No content sanitization is specified for the ingested files.
[SAFE]: No security issues were detected. The skill focuses on documentation management, and all file operations are localized to project-specific markdown files. No network operations, hardcoded credentials, or obfuscated contents were found.

domain-modeling