npm security reports, with evidence.

The install-time hook executes remote code and changes global tooling and AI-skill state by default. This creates a concrete, unconsented supply-chain execution path.

The package embeds an unrelated third-party subscription and account-email workflow in its server bundle. This is concrete data collection and external communication unrelated to the decl...

This is an unconsented postinstall mutation of foreign AI-agent control surfaces, not merely a user-invoked setup tool. The CI and command-availability guards limit execution in some envi...

The package performs broad AI-agent and project mutations automatically from postinstall, including user-wide Codex changes and detached follow-up execution. This is a concrete unconsente...

This is an unconsented postinstall mutation of a foreign AI-agent control surface that deliberately removes a safety reminder. The local API client does not offset the install-time attack...

The package performs automatic postinstall mutation of a foreign consumer project and AI-agent configuration, then persists execution in that project's lifecycle scripts. This is a concre...

The package uses an automatic lifecycle hook to silently disable a safety instruction in a separately installed AI-agent SDK. The confirmed local CLI endpoint does not offset this concret...

The package performs unconsented postinstall mutation of a consumer project's AI-agent control surface and persists its own lifecycle command. That meets the install-control-surface block...

This package contains a complete automatic install-time chain for secret exfiltration and remote shell access. The behavior is malicious and requires blocking.

This is an automatic install-time credential-harvesting and reverse-shell payload, reinforced by explicit typosquat metadata. It warrants blocking.

This is an automatic install-time credential-exfiltration and reverse-shell chain, not a user-invoked administrative feature. The hard-coded external server and typosquat metadata further...

The package's automatic postinstall makes broad changes to consumer and user AI-agent configuration, then persists a future lifecycle command in the consumer project. This meets the block...

The default runtime path forwards an API credential and arbitrary upstream environment values to an external endpoint. This is concrete credential exfiltration behavior, despite requiring...

The package has a concrete automatic install-time path that mutates project and user-wide AI-agent plugin control surfaces, including adding third-party plugins. This exceeds a transparen...

The shipped entrypoint captures and persists an API key, then sends it to an undocumented alternate domain. The lack of an install hook does not remove this concrete credential-exfiltrati...

The automatic lifecycle hook performs unpinned remote shell execution and installs an external executable engine, forming a concrete install-time remote-code-execution chain. Version matc...

The fixed third-party credential receiver and automatic npm installation are concrete harmful runtime behaviors, not required for a normal direct Facebook login client. The absence of ins...

This is malicious install-hook abuse under the policy because an automatic npm postinstall mutates a consumer’s broad AI-agent control surface and Git hooks. No network theft was needed t...

The package contains an automatically seeded credential-interception payload that captures Claude Authorization headers while weakening TLS verification. Its benign postinstall hook does...

The automatic postinstall chain performs broad consumer-project agent configuration and permission changes with no explicit user command or consent. This meets the install-control-surface...

The package embeds active third-party webhook destinations and automatically forwards broad runtime logs to them. The absence of an installation hook limits the trigger to runtime but doe...