graphhopper-selfhost

Pass

Audited by Gen Agent Trust Hub on Nov 6, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill directs the user to official GitHub repositories for the GraphHopper and OpenRailRouting projects to obtain executable JAR files or source code. These are well-known, legitimate sources for the tools described in the skill.
  • [COMMAND_EXECUTION]: Instructions include standard Java commands for data processing (import) and API hosting (serve). These commands use appropriate flags for memory management and configuration, and are necessary for the skill's primary function.
  • [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent on how to handle external OpenStreetMap data files (.pbf).
  • Ingestion points: OpenStreetMap .pbf data files (referenced in SKILL.md).
  • Boundary markers: Not applicable, as the data is binary and processed by external compiled Java tools.
  • Capability inventory: Execution of Java commands, file system writes for graph caching, and local network requests via curl for verification.
  • Sanitization: Data parsing and validation are handled internally by the GraphHopper/OpenRailRouting engines.
  • [SAFE_PRACTICE]: The skill explicitly warns against exposing the service to the public internet without user awareness by defaulting bind_host to localhost and requiring a deliberate change for LAN access.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 6, 2026, 04:41 PM