add-org
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes curl via the Bash tool to send a POST request to a LiteLLM proxy server for organization management.
- [INDIRECT_PROMPT_INJECTION]: The skill takes user-supplied parameters to construct an API request, creating a minor injection surface. 1. Ingestion points: User parameters for organization alias, models, and budget. 2. Boundary markers: Data is structured within a JSON request body. 3. Capability inventory: Shell command execution via curl. 4. Sanitization: Relies on the agent to format the JSON payload correctly.
- [SAFE]: The skill references official documentation and is authored by the service vendor, BerriAI, performing expected administrative actions.
Audit Metadata