azure-eventgrid-java

Warn

Audited by Gen Agent Trust Hub on Feb 14, 2026

Risk Level: MEDIUMPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [Indirect Prompt Injection] (MEDIUM): The skill provides patterns for parsing untrusted data from external sources (JSON payloads from webhooks) which can then influence agent behavior or be used in network operations. 1. Ingestion points: EventGridEvent.fromString(jsonPayload) and CloudEvent.fromString(cloudEventJson) in SKILL.md. 2. Boundary markers: Absent. No delimiters or instructions to ignore embedded commands are provided for the parsed content. 3. Capability inventory: EventGridPublisherClient.sendEvent() and EventGridReceiverClient.acknowledge() involve network operations and state modification across Azure services. 4. Sanitization: Absent. Code snippets demonstrate direct deserialization to objects without validation or sanitization of the input string.
  • [Unverifiable Dependencies & Remote Code Execution] (INFO): The skill references the Maven dependency com.azure:azure-messaging-eventgrid:4.27.0.
  • Source: Trusted (Microsoft/Azure).
  • Severity: Downgraded to INFO per [TRUST-SCOPE-RULE].
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 14, 2026, 04:59 PM