biz-health-check

Warn

Audited by Gen Agent Trust Hub on Aug 20, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill requires the use of the npx package runner to fetch and execute the @nomadamas/k-skill package from the NPM registry. This package provides the runtime instructions and execution environment.\n- [REMOTE_CODE_EXECUTION]: Executing remote packages via npx constitutes a remote code execution pattern. This tool is used to both fetch instructions and run the main script.\n- [DYNAMIC_EXECUTION]: In scripts/biz_health_check.py, the skill dynamically imports and executes Python modules from computed paths on the local file system using importlib.util. These modules correspond to sibling skills that provide specific data lookups.\n- [COMMAND_EXECUTION]: The skill instructions direct the agent to execute shell commands (e.g., npx ... exec) to perform its primary business logic.\n- [DATA_EXFILTRATION]: The skill sends business registration numbers and names to a configurable proxy server (--proxy-base-url) to interface with public APIs. This data flow is central to the skill's due diligence purpose.\n- [INDIRECT_PROMPT_INJECTION]:\n
  • Ingestion points: External data is ingested from multiple public APIs (National Tax Service, National Pension Service, etc.) through a proxy server.\n
  • Boundary markers: The skill does not implement explicit delimiters or warnings to ignore instructions that might be embedded in the retrieved API data.\n
  • Capability inventory: The skill has the ability to execute shell commands and dynamically load Python code.\n
  • Sanitization: Business numbers are validated via regex, but response content from the external APIs is processed without specific filtering or escaping.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Aug 20, 2026, 12:02 AM
Security Audit — agent-trust-hub — biz-health-check