earth2studio-install

Pass

Audited by Gen Agent Trust Hub on Jun 12, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [SAFE]: The skill is designed to guide users through the installation of Earth2Studio. It includes robust safety guidelines that explicitly forbid the agent from performing any package installations or system modifications without manual user execution, ensuring the user maintains control over their environment.
  • [EXTERNAL_DOWNLOADS]: The skill directs the agent to fetch up-to-date installation parameters from official NVIDIA documentation hosted on GitHub Pages (nvidia.github.io). This is a well-known and trusted source for software documentation, used here to ensure the accuracy of the commands provided to the user.
  • [COMMAND_EXECUTION]: The skill contains instructions for generating installation commands for tools like uv, pip, and system package managers (apt-get). These are presented as guidance for the user; the skill's core instructions prevent the agent from executing these commands itself.
  • [PROMPT_INJECTION]:
  • Ingestion points: The skill retrieves live content from https://nvidia.github.io/earth2studio/userguide/about/install.html (SKILL.md).
  • Boundary markers: No specific delimiters are defined for the documentation content in the prompt instructions.
  • Capability inventory: The agent is restricted to providing text-based instructions and is explicitly forbidden from running subprocesses, writing to the filesystem, or making unauthorized network calls.
  • Sanitization: The skill relies on the trust level of the source domain (NVIDIA) and does not specify internal sanitization for the fetched documentation content.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 12, 2026, 07:42 PM
Security Audit — agent-trust-hub — earth2studio-install