autopilot
Warn
Audited by Socket on Oct 22, 2026
1 alert found:
SecuritySecuritySKILL.md
MEDIUMSecurityMEDIUM
SKILL.md
SUSPICIOUS: the core behavior matches the stated purpose of automating commits, but the skill enables ongoing autonomous source-control actions with external effects, including pushes and PR submission, after a single invocation. No clear credential theft or hidden exfiltration appears, and the named mainstream CLIs are generally verifiable, but transitive skill use, unseen local scripts, and persistent no-confirmation operation make this a medium/high-risk workflow skill rather than a benign documentation helper.
Confidence: 88%Severity: 72%
Audit Metadata