npm AI-agent policy reports, with evidence.

The postinstall script writes command hooks into Claude Code's global settings file without asking the user. That is unconsented install-time mutation of a foreign AI-agent control surfac...

The postinstall hook mutates Claude Code and Codex MCP configuration without user consent, which is an unconsented install-time write to foreign AI-agent control surfaces. That is a block...

The automatic lifecycle hook persistently grants PATH-resolved commands execution in Claude Code's global hook configuration. This is concrete install-time AI-agent control-surface hijack...

This is an unconsented install-time mutation of a foreign AI-agent control surface. The lifecycle trigger and global hook writes establish a concrete control-hijack attack surface.

This is an unconsented postinstall mutation of a foreign, broad AI-agent control surface. The conditional executable lookup does not remove the automatic persistent configuration change w...

This is an unconsented install-time mutation of a broad, foreign AI-agent control surface. The conditional executable check does not remove the persistence or broad event scope.

This is an unconsented install-time mutation of a foreign AI-agent control surface, accompanied by persistent user-environment and launcher changes. The isolated virtual environment does...

Global npm install runs postinstall.js, which downloads a native binary without integrity checks and executes it with --configure-clients so it can write this package into other AI apps'...

The package performs unconsented postinstall mutation of broad AI-agent control surfaces and establishes persistence. This meets the install-control-surface blocking rule even though the...

This is a concrete unconsented postinstall chain that mutates broad AI-agent configuration surfaces and registers persistence. The loopback-only dashboard and lack of observed credential...

Global postinstall runs prumo install --all and mutates Claude, Codex, and Kiro control surfaces without an explicit user install command, which is unconsented foreign agent-config writes...

Global postinstall runs unattended prumo install --all, which writes this package's skill and always-on PO First instructions into Claude Code, Codex, and Kiro user configs and registers...

The automatic lifecycle hook performs concrete, broad mutation of the consumer's AI-agent workspace, including deletion. This meets the install-time agent-control-surface blocking policy.

npm postinstall mutates every detected coding-agent MCP config and then fetches an unsigned native zip from updater.kogiqa.com with no integrity check. That is unconsented foreign agent-c...

The postinstall script executes a native binary with --install-agent-hooks, and the package states that this writes Codex and Claude Code hooks without a separate user command. That is un...

The published postinstall mutates foreign AI-agent instruction files without consent, which is install-time control-surface hijack. The bundled CLI is a user-invoked wallet product and do...

The postinstall hook automatically runs connect install, which copies this package's skills into many third-party agent skill directories under the user home. That is unconsented install-...

The postinstall script silently writes Agentgates command instructions into Claude Code, Cursor, and Windsurf user config files, which is unconsented mutation of foreign AI-agent control...

The postinstall hook drives a third-party Chrome MCP bridge to register a native messaging host without a user command, which is unconsented mutation of a foreign AI-agent control surface...

Automatic postinstall writes a native binary and config into ~/.grok, a foreign Grok agent control surface, from a different npm scope. That unconsented install-time hijack is concrete at...

The install script launches a bundled binary and syncs hooks into Claude Code's user directories without a separate setup command. That is unconsented postinstall mutation of a foreign, m...