gmgn-token-buy
Pass
Audited by Gen Agent Trust Hub on Sep 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill identifies that on-chain token metadata (names, symbols, and descriptions) is attacker-controlled and could contain malicious instructions designed to bypass safety filters or influence agent behavior. It explicitly instructs the agent to treat these fields as data only and to respect sanitization flags from the CLI tool.
- [COMMAND_EXECUTION]: The skill relies on structured shell commands via the
gmgn-clibinary. It uses these commands to fetch market data, token security info, and gas prices. The skill includes logic to verify the CLI environment and correctly handle errors or missing dependencies before proceeding. - [EXTERNAL_DOWNLOADS]: The skill requires the
gmgn-clitool, which is an external dependency provided by the author (gmgnai). The instructions guide the user to install this tool from a standard registry (npm) if it is missing, which is consistent with the skill's primary purpose. - [DATA_EXFILTRATION]: The skill is designed to handle API keys securely by relying on the CLI's internal configuration management rather than storing or transmitting keys directly. It also explicitly avoids handling private keys, delegating signing operations to a separate execution skill.
Audit Metadata