expanso-priority-queues

Pass

Audited by Gen Agent Trust Hub on Jul 26, 2026

Risk Level: SAFEDATA_EXFILTRATION
Full Analysis
  • [DATA_EXFILTRATION]: The skill exposes a potential indirect prompt injection surface by ingesting external data via an HTTP server and processing it through logic that determines downstream routing.
  • Ingestion points: File pipeline.yaml configures an http_server input at 0.0.0.0:8080/events that accepts POST requests.
  • Boundary markers: None detected. The incoming JSON documents are parsed and mapped directly without delimiters or explicit instructions for the agent to ignore embedded commands.
  • Capability inventory: The skill uses expanso run (File run.sh) and routes data to multiple Kafka topics (File pipeline.yaml) based on calculated priority scores.
  • Sanitization: The mapping logic uses this.severity.or("INFO").uppercase() and uuid_v4() for normalization, but does not perform escaping or filtering of the raw content before processing or routing to downstream Kafka outputs.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 26, 2026, 05:44 AM
Security Audit — agent-trust-hub — expanso-priority-queues