confluent-tuning-advisor
Pass
Audited by Gen Agent Trust Hub on Sep 25, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill adheres to a strict advisory-only model, explicitly forbidding any write or mutation operations on the Kafka cluster, which ensures the human administrator remains in full control.
- [SAFE]: Security instructions explicitly prevent the agent from reading sensitive files like
.envand require the verification of environment variables without printing their values to the output. - [SAFE]: The skill includes comprehensive routing logic to decline and hand off out-of-scope tasks (e.g., Flink tuning, application code development, or WarpStream clusters), preventing unintended behavior or misconfigurations in unsupported environments.
- [SAFE]: All external resource references (documentation and APIs) target official and well-known Confluent domains, and the use of synthetic data in examples follows best practices for safety.
Audit Metadata