gmgn-market
Pass
Audited by Gen Agent Trust Hub on Nov 14, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill requires the global installation of the
gmgn-clipackage via NPM (npm install -g gmgn-cli). This is a vendor-provided tool used to interface with the GMGN API. - [COMMAND_EXECUTION]: Relies on shell commands to perform market analysis and system diagnostics.
- Executes
gmgn-clifor all market data retrieval and configuration. - Executes system utilities like
ifconfig,ip addr, andgrepto diagnose network interface settings. - [DATA_EXFILTRATION]: Accesses sensitive system metadata related to network configuration.
- The instructions command the agent to list all network interfaces and their IPv6 addresses using
ifconfig | grep inet6orip addr show | grep inet6. This exposes local network topology and internal interface details to the agent context. - Performs a network request to a well-known IP discovery service (
https://ipv6.icanhazip.com) to determine outbound IP version. - [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from various blockchains which acts as an attack surface for prompt injection.
- Ingestion points: Token names, symbols, contract descriptions, and social media metadata retrieved from the GMGN API via
gmgn-cli(SKILL.md). - Boundary markers: The skill does not define explicit delimiters or instructions to ignore embedded commands for the data returned from the API and displayed to the user.
- Capability inventory: While this specific file primarily focuses on data retrieval, it references external workflows for token research and automated swapping, which could be influenced by malicious metadata.
- Sanitization: The documentation states that
market searchfields are metadata-sanitized, but this sanitization is not explicitly confirmed for other endpoints liketrendingortrenchesfeeds.
Audit Metadata