The manifest has no install lifecycle hook, and the inspected high-risk source resolves to user-invoked usage requests, a local authenticated warm-check relay, DNS resolution, and a test-scoped fetch stub. These behaviors do not establish malicious activity or an unrelated credential recipient.
Static reason
High-risk behavior combination matched malicious policy.; previous stored version diff introduced dangerous source
Trigger
The usage feature runs when its command is invoked; the warm check runs only for a noninteractive check command.
Impact
No unrelated credential transfer, external command-output disclosure, or install-time attack was identified.
Mechanism
Usage credentials go to fixed provider endpoints with redirects rejected. Warm-check output is relayed through a local authenticated process channel, while the flagged test replaces fetch temporarily and restores it.
Network endpoints3
api.anthropic.com/api/oauth/usagechatgpt.com/backend-api/wham/usagecli-chat-proxy.grok.com/v1/billing?format=credits