gmgn-wallet-analysis
Pass
Audited by Gen Agent Trust Hub on Sep 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The analyze.py script uses subprocess.run() to invoke the gmgn-cli utility. The implementation follows security best practices by passing arguments as a list, avoiding shell execution, and performing format validation on wallet addresses before execution.
- [EXTERNAL_DOWNLOADS]: The skill depends on gmgn-cli, which is installed via npm. As a vendor-specific tool corresponding to the skill's author, this is a standard requirement for the skill's operation.
- [INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect injection as it processes externally-sourced token names and metadata tags. It addresses this risk with clear instructions to the agent to treat these fields strictly as data and to ignore any embedded instructions or claims of authority. Ingestion points: Activity logs, holdings data, and token metadata fetched from the GMGN platform via gmgn-cli. Boundary markers: Explicit instructions in SKILL.md direct the agent to disregard instructions within data fields. Capability inventory: Limited to read-only data retrieval through vendor-specific CLI tools; no general file-write or arbitrary code execution capabilities. Sanitization: The skill enforces address format checks for EVM and Solana chains to prevent malformed inputs from reaching the command-line interface.
Audit Metadata