building-streaming-pipelines

Pass

Audited by Gen Agent Trust Hub on Mar 13, 2026

Risk Level: SAFE
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides installation instructions for standard, well-known Python libraries from PyPI, including confluent-kafka, paho-mqtt, and nats-py.
  • [DATA_EXFILTRATION]: Network operations described in the documentation target local or well-known public test brokers (e.g., broker.emqx.io) and are essential for the primary functionality of a streaming data skill.
  • [PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection as it facilitates the ingestion of untrusted data from external streams (Kafka, MQTT, NATS).
  • Ingestion points: Consumer logic in references/kafka.md, references/mqtt.md, and references/nats.md reads message payloads from external brokers.
  • Boundary markers: The provided examples do not implement specific boundary markers or instruction-delimitation techniques.
  • Capability inventory: The skill includes network publication capabilities (producers) and standard file system operations if users extend the processing logic.
  • Sanitization: Code snippets demonstrate basic JSON parsing but do not explicitly detail sanitization or validation of untrusted payload content before further processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 13, 2026, 06:12 PM