npm AI-agent policy reports, with evidence.

The automatic postinstall hook performs forced, recursive writes to both Codex and Claude Code skill locations. That concrete unconsented mutation meets the install-time foreign AI-agent...

This is a concrete unconsented postinstall mutation of broad, foreign AI-agent control surfaces, combined with persistence and machine-identity cloud registration. The global/macOS guards...

This is concrete, unconsented postinstall mutation of foreign and broad AI-agent control surfaces, including permission settings. No network exfiltration is needed for the install-control...

This is an unconsented postinstall mutation of a foreign, global AI-agent control surface. The OAuth code is user-invoked and does not mitigate the install-time persistence.

Source directly confirms automatic, cross-agent skill installation during npm postinstall. This meets the blocking policy regardless of the absence of network activity in that lifecycle s...

Although no network exfiltration was found, the package concretely performs unconsented postinstall mutation of broad third-party AI-agent control surfaces. This meets the firewall block...

This is a concrete unconsented npm postinstall mutation of a foreign AI-agent control surface. The conditional hook/MCP path does not mitigate the unconditional skills write.

The package’s postinstall silently modifies global Claude Code configuration and installs broad lifecycle hooks, meeting the firewall block policy for unconsented AI-agent control-surface...

The lifecycle script performs a concrete default postinstall mutation of the consumer’s Cursor configuration and installs automatic hook execution. This meets the firewall block boundary...

The install hook performs unconsented mutation of both the consuming project's package manifest and GitHub Copilot instruction surface. That concrete lifecycle behavior meets the block bo...

The package has concrete install-time project and AI-agent control-surface writes, not merely dormant CLI functionality. This meets the block boundary for unconsented postinstall mutation...

The package performs unconsented postinstall writes to the consumer project and GitHub Copilot instruction surface. Although no install-time exfiltration was found, this meets the blockin...

The package performs a concrete, automatic postinstall write chain into the consuming project's AI-agent configuration rather than requiring an explicit user command. This meets the block...

Source inspection confirms concrete postinstall mutation of broad, foreign AI-agent plugin/control surfaces. This meets the firewall block policy regardless of the package's stated config...

The global-install guard limits the trigger but does not make the mutation user-invoked. The package performs the policy-defined blocking behavior even though no network exfiltration was...

The automatic multi-vendor global AI-agent skill installation is concrete unconsented control-surface mutation, meeting the block policy. The remote binary download increases impact but i...

The package performs a concrete global-install postinstall modification of Cursor configuration and commands. The global-install guard and lack of network activity do not remove the uncon...

This is concrete install-time mutation of a foreign AI-agent configuration, meeting the firewall block boundary despite its limited intended UI effect.

This is an unconsented postinstall mutation of the foreign Cursor AI-agent control surface. The global-install guard narrows exposure but does not provide user consent.

This is an unconsented postinstall mutation of a broad AI-agent control surface, not merely a package-local setup. The guarded detection does not obtain consent, and the same path also en...

The package’s postinstall path performs unconsented cross-vendor AI-agent configuration and persistence, which meets the blocking policy. The absence of a fixed exfiltration host does not...