hyperpod-version-checker

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The script scripts/hyperpod_check_versions.sh executes several system and diagnostics commands to collect version data. These include nvidia-smi, nvcc, mpirun, fi_info, lsmod, modinfo, docker, containerd, kubectl, and nvidia-ctk. These operations are consistent with the skill's purpose. \n- [INDIRECT_PROMPT_INJECTION]: The skill ingests data from the system environment and AWS services, creating a surface for indirect prompt injection.\n
  • Ingestion points: System command outputs, environment variables, and the AWS IMDS endpoint (169.254.169.254) in scripts/hyperpod_check_versions.sh.\n
  • Boundary markers: The script generates structured reports and uses jq to provide escaped JSON output, establishing data boundaries.\n
  • Capability inventory: The script uses curl for metadata access and various package/binary interrogation tools.\n
  • Sanitization: The script uses jq to ensure that gathered data is safely escaped before being presented in JSON format.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 05:58 PM
Security Audit — agent-trust-hub — hyperpod-version-checker