Security

The security boundary is mostly token handling, environment selection, and logging. Bot business code usually runs inside gateway event callbacks and then calls REST through the event session.

Token

Token holds the app_id and secret. Do not write the secret into the repository, logs, or panic output. Use safe_display() when you need a log-friendly representation:

tracing::info!("starting with {}", token.safe_display());

Prefer Token::from_env() for loading credentials:

• QQ_BOT_APP_ID
• QQ_BOT_SECRET

Token::validate() only checks that fields are non-empty. Real authentication results come from REST calls or gateway identify.

Sandbox and Production

The last argument to Client::new(token, intents, handler, is_sandbox) chooses sandbox or production. Use sandbox during development and validation, then switch to production after event, permission, and message flows are confirmed.

When constructing HttpClient independently, pass the same sandbox flag:

let http = HttpClient::new(30, true)?;

Logging

The gateway and HTTP layers emit diagnostics through tracing. In production, avoid broad debug logging that collects raw request bodies into external systems, especially logs containing user message content, openids, or token-related errors.

See Also

• Token
• Configuration
• Gateway