geobase-worker-overture-import
Pass
Audited by Gen Agent Trust Hub on Sep 28, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill implements secure secret management by explicitly instructing the agent not to handle API keys directly and advising users to use gitignored environment files.
- [SAFE]: All network operations and resource references target vendor-controlled infrastructure or well-known mapping documentation sites.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted spatial data (GeoJSON) and user-defined metadata which are processed by a remote worker.
- Ingestion points: aoiGeoJson, areaName, and tablePrefix fields in SKILL.md.
- Boundary markers: The protocol uses structured JSON payloads, though explicit boundary warnings for the LLM are absent.
- Capability inventory: The skill triggers HTTP POST requests to worker API endpoints to initiate data processing jobs.
- Sanitization: The instructions mandate strict validation of table prefixes via regular expressions and GeoJSON coordinate verification before submission.
Audit Metadata