npm AI-agent policy reports, with evidence.

The package performs unconsented postinstall mutation of broad foreign AI-agent control surfaces. This meets the firewall block boundary even though no install-time exfiltration was found.

Source confirms an unconsented postinstall mutation of multiple foreign AI-agent control surfaces. This meets the firewall block policy regardless of the installer’s lack of network activ...

This is unconsented postinstall mutation of foreign/broad AI-agent environments, coupled with remote native payload installation. The source establishes the behavior directly.

Source confirms an unconsented install-time mutation of a broad, foreign AI-agent control surface. Per policy, this is a publish-block condition despite the absence of credential exfiltra...

This is unconsented postinstall mutation of foreign AI-agent control surfaces, meeting the block boundary. No exfiltration or remote payload is needed for this concrete control-surface at...

Source confirms an unconsented postinstall mutation of ~/.agents/skills, a foreign/broad AI-agent control surface. This meets the blocking policy regardless of the unrelated CLI and Git-h...

This is concrete unconsented postinstall mutation of foreign AI-agent control surfaces, combined with remote executable retrieval and host-level configuration changes. The global-install...

This is concrete unconsented postinstall mutation of foreign AI-agent control surfaces, compounded by unauthenticated remote binary execution and host configuration changes.

This is concrete, unconsented postinstall mutation of multiple foreign AI-agent control surfaces, meeting the blocking policy. The absence of a hard-coded exfiltration host does not remov...

This is a concrete postinstall-time mutation of foreign/broad AI-agent control surfaces, including automatic Codex plugin enablement. The absence of install-time networking does not remov...

The lifecycle hook performs active AI-agent configuration and hook installation in the consumer project without an explicit user command. This meets the blocking policy even though no net...

This is unconsented install-time destructive mutation of an AI CLI credential/control surface coupled with automatic account provisioning. It exceeds a safe legacy migration or ordinary C...

Source confirms an unconsented postinstall mutation of multiple foreign/broad AI-agent control surfaces, which meets the blocking policy. The package-aligned skill content and checksum ch...

The source confirms a concrete postinstall mutation of several foreign/broad agent control surfaces, meeting the block policy. The benign-looking skill content and checksum check do not r...

The installer directly performs broad, persistent global agent-control-surface writes during postinstall. User-invoked network/credential tooling does not negate that install-time behavior.

This is concrete unconsented postinstall mutation of a shared/foreign AI-agent credential surface, not a user-invoked migration. The competing-service marker deletion and credential repla...

This is concrete unconsented postinstall mutation of foreign/broad AI-agent control surfaces, meeting the blocking policy. Optional local-only MCP transport does not mitigate the install-...

The lifecycle script performs broad cross-vendor agent configuration and MCP registration without explicit user setup. Local-only networking and the apparent browser-automation purpose do...

The lifecycle hook performs persistent global Claude Code command installation, which meets the firewall block boundary for unconsented postinstall mutation of a foreign/broad AI-agent co...

The package performs unconsented postinstall mutation of a foreign AI-agent configuration and enables external agent control channels. This meets the firewall blocking boundary even thoug...

This is a concrete unconsented install-time mutation of broad foreign AI-agent control surfaces. Guardrails and lack of observed exfiltration do not remove that attack surface.