The manifest exposes a Great Arrow CLI, and the flagged hook is opt-in: it sends the prompt with a Great Arrow token to the Great Arrow service for tool suggestions. Targeted inspection found no concrete malicious credential flow, metadata collection, or concealed payload execution.
Static reason
High-risk behavior combination matched malicious policy.
Trigger
The prompt suggestion hook runs only after the user enables it.
Impact
Great Arrow receives the prompt for tool suggestions, as described by the feature; no unrelated credential theft or other attack behavior was identified.
Mechanism
When enabled and configured with a Great Arrow token, the hook posts the prompt to the Great Arrow tool suggestion endpoint.
Evidence
dist/hooks/gad-suggest-tools.mjs~/.great-arrow/hook-config.json
Network endpoints1
www.greatarrow.ai/api/mcp/tools-suggest