inworld
Pass
Audited by Gen Agent Trust Hub on Sep 22, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: Comprehensive analysis of the instructions and code examples revealed no malicious patterns, prompt injections, or unauthorized data exfiltration techniques.
- [DATA_EXPOSURE]: The skill correctly implements credential management by using environment variables (e.g.,
os.getenv('INWORLD_API_KEY')) instead of hardcoded secrets in the provided Python examples. - [COMMAND_EXECUTION]: The provided code snippets perform expected operations, such as making HTTP POST requests to the official Inworld API and writing generated audio to local files.
- [OBFUSCATION]: The use of
base64.b64decodeis appropriate for its context, handling binary audio data returned as a string from the API, and does not hide malicious logic.
Audit Metadata