arcgis-ai-components
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill documentation recommends loading the ArcGIS Maps SDK and its associated AI components directly from the official Esri CDN at
js.arcgis.com. This is a standard and safe integration pattern for this service. - [INDIRECT_PROMPT_INJECTION]: The skill enables a natural language chat interface that processes user-supplied text to perform actions on map data and execute specialized agents.
- Ingestion points: User messages submitted via the
arcgis-assistantcomponent (SKILL.md) and descriptive data ingested from map layer metadata. - Capability inventory: The agents can execute map navigation commands, perform data analysis/summarization, and invoke custom tools such as network requests (
fetch) defined in custom agent registrations. - Boundary markers: The architecture uses a routing system to dispatch messages to specific agents (navigation, exploration, help), providing an inherent structural scope to the agent's actions.
- Sanitization: The documentation emphasizes the importance of high-quality metadata and embeddings to ensure agent accuracy, though it relies on the underlying ArcGIS platform for primary input sanitization.
Audit Metadata