The Agent Skills Directory

[SAFE]: No malicious patterns or security risks were identified. The skill's primary functions involve reading local configuration files and organizing items into directories, both of which are common and appropriate behaviors for an AI agent performing organizational tasks. The architecture includes specific 'Anti-Patterns' and 'Success Criteria' that enforce user oversight and dynamic discovery over hardcoded paths.
[PROMPT_INJECTION]: Indirect Prompt Injection Surface: The skill ingests data from external CLAUDE.md files located within the /Brains/ directory to build its routing index. This creates a surface where malicious instructions embedded in those configuration files could attempt to bias the agent's routing decisions or manipulate its categorization logic. This risk is effectively neutralized by the skill's 'Sequential Questioning Protocol' and confidence thresholds which require explicit human confirmation for any non-obvious matches.
Ingestion points: File content parsed from /Brains/*/CLAUDE.md.
Boundary markers: None (the parsing logic does not specify the use of delimiters or 'ignore' instructions for the content of the configuration files).
Capability inventory: The skill utilizes file system operations including directory listing, file reading, and the creation of new directories and files ('New Brain Creation') through user-guided prompts.
Sanitization: No content validation or sanitization is mentioned for the text extracted from the configuration files.

routing-brains