physicsnemo-discover

Pass

Audited by Gen Agent Trust Hub on Jul 3, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill performs a shallow clone of the official NVIDIA PhysicsNeMo repository to access up-to-date files and documentation for discovery.
  • Evidence: SKILL.md contains instructions to execute "git clone --depth 1 https://github.com/NVIDIA/physicsnemo" into a temporary directory.
  • [COMMAND_EXECUTION]: The skill utilizes shell commands to manage the repository clone and verify file paths before presenting them to the user.
  • Evidence: Uses "git clone" for retrieval and "Bash ls -d" for path verification as defined in SKILL.md and references/RECIPES.md.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it reads and summarizes content from an external repository that could potentially contain malicious instructions.
  • Ingestion points: Repository content read via Read, Glob, and Grep tools as specified in references/RECIPES.md.
  • Boundary markers: The skill includes explicit instructions to treat the cloned repository as read-only and never execute or import its contents.
  • Capability inventory: Tools include git, ls, Read, Glob, and Grep for file and directory analysis.
  • Sanitization: The skill does not implement explicit sanitization or filtering of the content read from the repository files before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 3, 2026, 11:06 AM
Security Audit — agent-trust-hub — physicsnemo-discover