vastai

Pass

Audited by Gen Agent Trust Hub on Aug 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the vastai CLI tool to manage GPU lifecycles, billing, and instance configuration. It provides explicit constraints (Rule 7) to prevent the agent from attempting to use unauthorized binaries or direct API calls, which serves as a security guardrail.
  • [EXTERNAL_DOWNLOADS]: References the installation of the vastai package from PyPI. As this package is the primary tool for the skill and belongs to the vendor (vast-ai), this is a standard and expected dependency.
  • [DATA_EXPOSURE]: The skill instructs the agent to read the user's SSH public key (~/.ssh/id_ed25519.pub) for the purpose of registering it with the remote host. This is a primary function of the skill and follows standard practices for the service.
  • [DYNAMIC_CONTEXT_INJECTION]: The documentation mentions the use of the ! shell execution prefix for 2FA login commands. This is used as a privacy recommendation to ensure that sensitive short-lived codes do not persist in the conversation history, rather than as a malicious injection pattern.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 12, 2026, 01:01 PM
Security Audit — agent-trust-hub — vastai