french-site-check

Pass

Audited by Gen Agent Trust Hub on Oct 11, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill requires a connection to a Model Context Protocol (MCP) server at 'https://mcp.terra-lab.ai/mcp'. This is a resource owned by the skill's author (terralabai) used to provide the underlying parcel and elevation data.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data returned from external tools (cadastral parcels and urban planning zones).
  • Ingestion points: Data enters the context via the french_parcel and get_elevation tool outputs (SKILL.md).
  • Boundary markers: No explicit delimiters or instructions to ignore embedded instructions are defined for the tool output.
  • Capability inventory: The skill is limited to making specific tool calls and providing download links; it does not have file-system write or shell execution capabilities.
  • Sanitization: The instructions direct the agent to report specific fields and provide returned links, which serves as a functional filter for the output data.
  • [COMMAND_EXECUTION]: No evidence of arbitrary shell command execution or unsafe subprocess spawning was detected.
  • [CREDENTIALS_UNSAFE]: No hardcoded credentials or sensitive environment variable access was found.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 11, 2026, 09:07 AM