vastai-sdk Audit report.

Warnings

• Dependency or permission surface needs review
• Permission surface may require sandboxing
• SKILL.md grants access to the entire vastai:* Python namespace, including destructive and costly operations such as destroy_instance, delete_deployment, unlist_machine, create_instance, and cloud_copy, without explicit confirmation or safety warnings.
• The API key can be passed as a literal in code examples, which risks secret leakage in agent logs or chat history; no guidance on key scoping or rotation is provided.
• Safe operating boundaries are not clearly defined, such as requiring user confirmation before irreversible actions or avoiding /root and / as copy destinations.
• The provided SKILL.md appears truncated/incomplete at the SSH Keys section, so some documented functionality is not fully covered.
• Quality score needs review
• Permission surface needs review: secrets or environment access, shell or command execution
• Dependency/runtime risk: command execution surface, credential or environment access
• Permission surface: secrets or environment access, shell or command execution

Method

This report combines public metadata, AI review output, repository freshness, install readiness, OpenAgentSkill events, quality scoring, trust checks, and the agent safety gate. It is not a full source-code security review.