together-code-interpreter
Pass
Audited by Gen Agent Trust Hub on Mar 30, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious behavior or security violations detected. The skill follows best practices for API integration and relies on official Together AI resources.
- [COMMAND_EXECUTION]: Implements remote Python execution in a managed sandbox as its core functionality, allowing for data analysis and stateful notebook-like workflows.
- [EXTERNAL_DOWNLOADS]: Fetches dependencies and interacts with APIs from official, well-known Together AI repositories and registries.
- [PROMPT_INJECTION]: Analyzed for indirect prompt injection risk due to the processing of user-provided code. Ingestion points: 'code' and 'files' fields in execution scripts. Boundary markers: None explicitly defined. Capability inventory: Remote Python runtime with access to common data science libraries. Sanitization: Execution is restricted to a managed sandbox environment controlled by the service provider.
Audit Metadata