streaming-data-skill
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- SAFE (SAFE): No malicious patterns, prompt injections, or persistence mechanisms were found. The skill serves as an educational and implementation guide for streaming data systems.- CREDENTIALS_UNSAFE (SAFE): All code examples for Kafka, Snowflake, and BigQuery ingestion use variable placeholders (e.g., 'acct', 'usr', 'key', 'localhost:9092') rather than hardcoded credentials.- PROMPT_INJECTION (SAFE): The SKILL.md and README.md files contain natural instructional language for system guidance without any attempts to bypass model safety filters or override system prompts.- EXTERNAL_DOWNLOADS (SAFE): The installation instructions use a standard 'git clone' for the skill repository itself, which is expected behavior for skill deployment.
Audit Metadata