venice-router
Pass
Audited by Gen Agent Trust Hub on Jul 17, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill description utilizes terminology commonly associated with safety bypass and jailbreaking. Phrases such as "uncensored mode", "no content filters", and "no refusals" are used to describe the Venice.ai platform. While consistent with the platform's features, these are high-risk patterns for prompt injection.\n- [COMMAND_EXECUTION]: The skill facilitates the execution of a local Python script (
venice-router.py) for its core operations, including model classification and routing.\n- [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface due to the handling of untrusted data.\n - Ingestion points: External data enters the agent context through the
--prompt,--conversation, and--toolscommand-line arguments in SKILL.md.\n - Boundary markers: The documentation lacks explicit boundary markers or instructions to ignore instructions embedded within the processed data.\n
- Capability inventory: The skill executes a local Python script that performs network operations (API calls) and cost management functions.\n
- Sanitization: There is no evidence of sanitization or validation of external inputs within the provided skill instructions.
Audit Metadata