truefoundry-mcp-servers

Pass

Audited by Gen Agent Trust Hub on Apr 15, 2026

Risk Level: SAFE
Full Analysis
  • [PROMPT_INJECTION]: The skill instructions include explicit security policies and defensive prompts that instruct the agent to reject raw credentials and require human-in-the-loop confirmation for deletions and external URL usage. These measures are designed to enhance the safety of the agent's operations.
  • [DATA_EXFILTRATION]: The scripts/tfy-api.sh helper script reads API keys and configuration from environment variables or a local .env file for authentication. This data is used solely to interact with the user-configured TrueFoundry platform endpoint and does not involve unauthorized data transmission to third-party domains.
  • [COMMAND_EXECUTION]: The skill utilizes the bash environment to execute the tfy-api.sh script (a wrapper for curl) and the tfy CLI. These commands are necessary for the skill's primary function of managing cloud resources and follow standard operational patterns for TrueFoundry tools.
  • [EXTERNAL_DOWNLOADS]: The skill documentation references pinned container images and packages from well-known registries such as AWS Public ECR, GitHub Container Registry (HuggingFace), and NVIDIA GPU Cloud (NGC). These references are part of standard configuration guidelines and do not involve unverified remote code execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 15, 2026, 06:05 AM