muapi-models
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONPERSISTENCE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill recommends installing the
muapi-clitool viapipornpmto facilitate file uploads and model execution. - [COMMAND_EXECUTION]: Shell commands such as
curlandcurl.exeare used to interact with the Muapi API and download generated files to the local project directory. - [INDIRECT_PROMPT_INJECTION]: The skill processes data and media URLs from external model results. While this is a standard integration workflow, it involves downloading untrusted content to the local system.
- Ingestion points: Predictions result endpoint (
/api/v1/predictions/{request_id}/result) inSKILL.md. - Boundary markers: None specified for model-generated content.
- Capability inventory:
curl(network/file write) andmuapi-cli(network access). - Sanitization: No specific sanitization process is described for model outputs before they are saved locally.
- [PERSISTENCE]: The skill provides instructions for using the
setxcommand on Windows to store theMUAPI_API_KEYenvironment variable across sessions, which is standard configuration for such integrations.
Audit Metadata