venice-ai-api
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill successfully documents the Venice.ai API without introducing malicious patterns or security vulnerabilities. It focuses on providing valid integration patterns for developers.
- [DATA_EXFILTRATION]: The instructions explicitly caution against hardcoding credentials and provide safe examples using environment variables (e.g.,
VENICE_API_KEY) for API key management. It emphasizes that keys should be server-side and never exposed to client-side code. - [INDIRECT_PROMPT_INJECTION]: The documentation identifies untrusted data sources such as web search results, scraped content, and file uploads. It provides specific remediation patterns, including the use of boundary markers (e.g.,
<document>tags) and explicit system instructions to treat external content as data rather than instructions. - [COMMAND_EXECUTION]: The included Python discovery script (
scripts/discover_models.py) uses standard library functions to fetch model catalog information from official API endpoints. It does not perform dangerous shell execution or incorporate untrusted remote payloads into execution flows.
Audit Metadata