The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the Python subprocess module to run gmgn-cli for data retrieval and performs network diagnostics using ifconfig and ip addr. The implementation follows secure practices by passing arguments as a list rather than a shell string, mitigating potential command injection vulnerabilities.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external blockchain data (token names, wallet tags) which presents an indirect injection surface.
Ingestion points: Data is ingested into the agent context through the outputs of the analyze.py script and the gmgn-cli tool.
Boundary markers: SKILL.md provides clear instructions to the AI agent to treat all third-party string fields as data only and to ignore any instructions embedded within them.
Capability inventory: The skill has capabilities to execute local CLI tools and provide detailed textual analysis.
Sanitization: The skill relies on natural language boundary markers and agent-level instructions to sanitize potentially malicious metadata.
[EXTERNAL_DOWNLOADS]: The skill requires the installation of the gmgn-cli utility via npm. Since this is the official tool provided by the vendor gmgnai, it is documented as a standard functional dependency.

gmgn-wallet-analysis