fleet-ops
Pass
Audited by Gen Agent Trust Hub on Jul 15, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill utilizes local CLI tools (
joelclaw fleet status,joelclaw fleet diff) to inspect system state. These operations are described as read-only and do not involve installing packages or altering services. - [SAFE]: The instructions include strong security guardrails, explicitly directing the agent to avoid copying credentials, publishing private manifests, or exposing network topology.
- [SAFE]: The skill mandates a human-review process for any fleet changes, explicitly stating that automatic 'apply' or 'plan' mechanisms do not exist within the skill context to prevent unauthorized mutations.
Audit Metadata