The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the project's codebase, which could potentially contain malicious instructions designed to influence the agent's behavior. * Ingestion points: Reads 'CONTEXT.md', 'CONTEXT-MAP.md', architecture decision records (ADRs) in 'docs/adr/', and project source code. * Boundary markers: None identified. The skill reads the files directly without specific delimiters or instructions to ignore embedded commands. * Capability inventory: The skill has the ability to read and update project documentation files ('CONTEXT.md', ADRs). It does not exhibit network access or arbitrary command execution capabilities. * Sanitization: No explicit sanitization or validation of the content read from the files is performed before processing.

domain-modeling