truefoundry-jobs

Pass

Audited by Gen Agent Trust Hub on May 8, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: Executes the tfy CLI and a provided bash helper script (tfy-api.sh) to perform job deployment and monitoring tasks. This involves creating YAML manifests and triggering runs through the TrueFoundry platform. It also includes a version detection script (tfy-version.sh) that checks local environment compatibility.
  • [EXTERNAL_DOWNLOADS]: Automatically installs the truefoundry Python package from PyPI if the CLI is missing. It also references and pulls container images from trusted registries including Amazon ECR, GitHub Container Registry, and NVIDIA Container Toolkit. These downloads are associated with official vendor tooling and registries.
  • [CREDENTIALS_UNSAFE]: Handles authentication by reading TFY_API_KEY and TFY_BASE_URL from environment variables or local .env files. Instructions mandate the use of the tfy-secret:// protocol for application-level secrets to ensure sensitive data is not exposed in logs or chat history.
  • [PROMPT_INJECTION]: Proactively mitigates indirect prompt injection risks by explicitly instructing the agent not to fetch or parse external release pages from third-party sites like GitHub or HuggingFace, relying instead on local pinned versions for verified image tags.
Audit Metadata
Risk Level
SAFE
Analyzed
May 8, 2026, 02:43 AM