The Agent Skills Directory

[SAFE]: The skill contains logic-based instructions for task classification and resource allocation. No malicious patterns, obfuscation, or unauthorized access attempts were detected.
[COMMAND_EXECUTION]: The skill specifies the use of a shell tool within its metadata and instructions. This is used for repository navigation and task execution as part of its standard orchestration role.
[PROMPT_INJECTION]: The skill defines an attack surface for indirect prompt injection by instructing the agent to process untrusted data sources.
Ingestion points: User goals, code blocks, logs, and external URLs (referenced in the 'Inputs To Assess' and 'Routing Procedure' sections of SKILL.md).
Boundary markers: None explicitly defined within the routing logic.
Capability inventory: Access to the shell and project-docs tools for file system interaction and command execution.
Sanitization: No specific sanitization or filtering logic is provided for the ingested data.

hermes-route