skills/holon-run/uxc/dune-mcp-skill/Gen Agent Trust Hub

dune-mcp-skill

Pass

Audited by Gen Agent Trust Hub on Mar 8, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to use uxc and a linked dune-mcp-cli command. These commands are used for legitimate interaction with the documented Dune API endpoint (api.dune.com).
  • [CREDENTIALS_UNSAFE]: The skill provides clear guidance on secure authentication using environment variables (DUNE_API_KEY) or secret managers like 1Password (op://). It correctly uses placeholders like {{secret}} instead of hardcoding any actual keys.
  • [EXTERNAL_DOWNLOADS]: The skill references the official Dune API endpoint (https://api.dune.com/mcp/v1). These references are necessary for the skill's primary function and target a well-known service provider.
  • [REMOTE_CODE_EXECUTION]: No patterns of remote script execution or dynamic code evaluation were detected. All execution is limited to pre-defined CLI operations.
  • [DATA_EXFILTRATION]: No evidence of sensitive file access or unauthorized data transmission was found. Network activity is confined to the official API domain for the purpose of retrieving blockchain data.
  • [PROMPT_INJECTION]: No attempts to override system prompts, bypass safety filters, or extract underlying instructions were found in the skill body or metadata.
  • [SAFE]: The validation script (scripts/validate.sh) is a standard utility for ensuring project structure and documentation consistency. It does not perform any dangerous operations.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 8, 2026, 09:42 AM