streaming-data

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • SAFE (SAFE): The provided TypeScript and Python code examples perform standard Kafka operations (connecting, producing, consuming) without any suspicious activity or attempts to access sensitive local files.
  • CREDENTIALS_UNSAFE (SAFE): References to 'password' within the documentation files ('cdc-patterns.md') are clearly identifiable as documentation placeholders for local development examples and do not constitute a hardcoded credential risk.
  • EXTERNAL_DOWNLOADS (SAFE): The skill references reputable and industry-standard libraries such as 'kafkajs' for Node.js and 'confluent-kafka' for Python. No untrusted or malicious dependencies were identified.
  • PROMPT_INJECTION (SAFE): The content consists primarily of technical documentation and code snippets; no instructions were found that attempt to override agent safety filters or system prompts.
  • INDIRECT_PROMPT_INJECTION (SAFE): The consumer examples ingest external message data but lack any dangerous capabilities (e.g., shell execution, eval) that could be exploited through the data stream.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:12 PM