soia-env-network-diagnose
Pass
Audited by Gen Agent Trust Hub on Aug 6, 2026
Risk Level: SAFE
Full Analysis
- [DATA_EXFILTRATION]: The Python probing script includes explicit sanitization logic to remove usernames, passwords, query strings, and fragments from URLs before reporting outcomes. This prevents the accidental exposure of sensitive credentials that might be included in provided URLs.
- [EXTERNAL_DOWNLOADS]: Diagnostic connectivity checks are performed against official vendor sites (such as Node.js and Python) and recognized public mirrors (including Tencent Cloud and Alibaba Cloud). These are recognized as safe endpoints for software development and troubleshooting.
- [COMMAND_EXECUTION]: The skill executes a local Python script for connectivity testing. The script is constrained to HEAD requests and relies solely on the Python standard library. Instructions explicitly forbid high-risk behaviors like piped shell execution or automated changes to system configurations.
- [SAFE]: The skill maintains a strict read-only posture, clearly defining its boundaries by refusing to modify system files, DNS settings, or proxy configurations. It prioritizes user authorization and provides transparent results through a desensitized reporting format.
Audit Metadata