gpt-image-2-exact-text-image
Pass
Audited by Gen Agent Trust Hub on Aug 12, 2026
Risk Level: SAFE
Full Analysis
- [CREDENTIALS_UNSAFE]: The skill securely manages API keys by storing them in a local configuration file (
~/.ai-hive/config.json) and enforcing restricted file permissions (0600) to prevent unauthorized access by other local users. - [DATA_EXFILTRATION]: Network communication is limited to the official vendor API endpoint (
ai-hive.iclip.cn) for task submission, polling, and media management. No unauthorized data exfiltration was detected. - [EXTERNAL_DOWNLOADS]: The skill facilitates the download of generated media assets from the service provider to the local machine as part of its core functionality.
- [COMMAND_EXECUTION]: The included Python script uses standard libraries and the
requestspackage to interact with remote services without employing unsafe execution patterns likeeval()or direct shell piping.
Audit Metadata