The Agent Skills Directory

[SAFE]: The skill provides logic for prioritizing and routing tasks to other agents based on attribute data stored in a local JSON file (.claude/attributes.json). No suspicious network operations or unauthorized file access patterns were identified.
[SAFE]: Use of the Agent tool is restricted to spawning specialized sub-agents (e.g., security-audit, test, refactor) as part of its documented routing functionality.
[SAFE]: The skill handles 'findings' as data to be processed, which is a standard pattern for orchestration skills. It does not contain instructions that attempt to bypass safety filters or ignore previous instructions.

attributes-route