The Agent Skills Directory

[SAFE]: Task Routing Logic
The skill defines a clear decision matrix for engine and model selection based on task signals like complexity and risk, ensuring appropriate engine use without bypassing security protocols.
[SAFE]: Data Context Sharing
Access to ~/.claude and ~/.qwen directories is utilized solely for sharing context artifacts, memory, and logs between the routed AI engines to ensure continuity.
[SAFE]: Command Orchestration
The skill references legitimate CLI commands such as qwen-dispatch and qwen-escalate to delegate tasks, which is consistent with its intended function as a unified router.

unified-router