The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local utility scripts, get-token.sh for Linux/macOS and get-token.ps1 for Windows, to fetch authentication tokens for every API request.
[SAFE]: The authentication process fetches tokens from an external service hosted on a well-known domain (jprx.m.qq.com). This is documented as a standard part of the skill's integration workflow.
[SAFE]: The skill includes robust safety rules (AI behavior contracts) specifically designed to prevent the leakage of access tokens. It explicitly forbids the AI from displaying, echoing, or discussing token values in any part of the conversation.

notion