vastai-cost-tuning
Pass
Audited by Gen Agent Trust Hub on Mar 4, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill consists of markdown documentation and code templates intended for user reference. No security risks were identified.
- [PROMPT_INJECTION]: The instructions are clear, task-oriented, and contain no attempts to override agent behavior, extract system prompts, or bypass safety guardrails.
- [DATA_EXFILTRATION]: There are no hardcoded credentials, sensitive file access patterns, or unauthorized network requests. All referenced domains (vastai.com) are official service providers related to the skill's purpose.
- [REMOTE_CODE_EXECUTION]: No remote code execution or untrusted dependency installations were found. The provided TypeScript and SQL snippets are static examples and are not executed dynamically by the agent.
- [COMMAND_EXECUTION]: The skill only requests permission for 'Read' and 'Grep' tools, which are limited to local file analysis and pose no risk of privilege escalation or system modification.
Audit Metadata