The flagged credential scanning and encrypted-payload handling belong to the CLI’s explicit key-rotation workflow, which uses browser approval and a device key. The prompt hook is opt-in, and telemetry is consent-gated; the reviewed source does not show an automatic exfiltration or unrelated payload execution path.
Static reason
High-risk behavior combination matched malicious policy.
Trigger
Key inventory and updates run through explicit `gad keys` commands. The prompt hook runs only after the user enables it; telemetry requires consent.
Impact
No automatic credential theft, unrelated data collection, or concealed payload execution was identified.
Mechanism
Enrollment reports environment-file key names and paths. Applying an update fetches a sealed payload and opens it with the enrolled device key. The opt-in prompt hook sends prompt text to the Great Arrow service.
Evidence
dist/index.jsdist/hooks/gad-suggest-tools.mjspackage.json