truefoundry-workspaces
Pass
Audited by Gen Agent Trust Hub on Apr 19, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is well-structured and follows platform security guidelines. It includes explicit instructions for the agent to use secret management (tfy-secret://) for sensitive credentials.
- [COMMAND_EXECUTION]: The script
scripts/tfy-api.shexecutescurlto communicate with the TrueFoundry API. The script includes input validation that blocks path traversal attempts (..) and restricts execution to a whitelist of standard HTTP methods. - [EXTERNAL_DOWNLOADS]: The skill references container images from well-known and trusted registries, including Amazon ECR (
public.ecr.aws/truefoundrycloud), GitHub Container Registry (ghcr.io/huggingface), and NVIDIA Container Registry (nvcr.io/nim). It also provides instructions for installing the official vendor CLI from PyPI. - [PROMPT_INJECTION]: The skill contains defensive instructions in
references/container-versions.mdthat explicitly forbid the agent from fetching or parsing content from external release pages, specifically to mitigate the risk of indirect prompt injection from third-party sources.
Audit Metadata