truefoundry-ml-repos
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes a local helper script located at
scripts/tfy-api.shto perform authenticated HTTP requests. This script serves as a wrapper forcurl, handling the injection of theTFY_API_KEYinto request headers and implementing basic safeguards against path traversal. - [EXTERNAL_DOWNLOADS]: The skill provides instructions in
references/prerequisites.mdfor the automated installation of thetruefoundryCLI package from the standard Python package registry (PyPI). This is a known vendor-provided tool required for deployment workflows. - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external TrueFoundry API endpoints (listing ML repositories and models). This metadata could potentially be manipulated by a malicious actor to include adversarial instructions intended to influence the agent's behavior.
- Ingestion points: API response data from TrueFoundry endpoints (e.g.,
/api/ml/v1/ml-repos) processed viascripts/tfy-api.sh. - Boundary markers: None explicitly implemented in the skill's data processing instructions.
- Capability inventory: The skill has the capability to execute bash scripts and interact with cloud deployment services.
- Sanitization: The skill relies on the agent's internal capability to safely parse JSON formatted data from the API responses.
Audit Metadata