The metadata says there are no keys/environment variables, but the README explicitly states it auto-fills `Authorization: QQBot {token}`, indicating platform credentials are in fact involved, likely injected by the host rather than manually entered by the user. This creates the normal risk of credential-backed administrative actions, but there is no explicit red flag showing credential exfiltration.

The README explicitly identifies the remote endpoint as `https://api.sgroup.qq.com` and uses HTTP requests against the QQ Open Platform for query and management operations. The network egress target is consistent with the stated functionality, with no evidence of data being sent to unrelated or unknown third-party endpoints.

The objective checks mark it as `prompt-only`, and the provided materials do not show any local process spawning, script execution, or invocation of additional system commands. Based on the available information, it does not appear to have local code execution capability.

This skill can read platform data such as guilds, channels, members, and permissions, and can also publish posts, manage announcements and schedules, and modify/delete channels, which amounts to read/write administrative access over remote service resources. There is no indication that it reads local files or requests unrelated excessive system permissions.

The source is a GitHub open-source repository, and the system reports `open-source` with an extremely strong community-adoption signal (about 377k stars), both of which materially reduce supply-chain concern. Although the license is not stated and maintenance status is unknown, the available materials do not show high-risk red flags such as closed source, abandoned opaque distribution, or suspicious delivery behavior.