The inspected entry points, lifecycle scripts, and flagged paths show ordinary build, scanning, runtime setup, and provider-call behavior. Credential routing and remote A2A requests include safeguards, and no unrelated collection or hidden attack behavior was identified.
Static reason
High-risk behavior combination matched malicious policy.; previous stored version diff introduced dangerous source
Trigger
Network and credential handling occurs when users run agent or provider features.
Impact
Expected requests to configured providers; no unrelated credential collection, concealed execution, or malicious install behavior was found.
Mechanism
Provider credentials are used for configured service calls; A2A calls use guarded transport, disable redirects, and cap response sizes.
Network endpoints2
api.anthropic.comapi.openai.com