The postinstall hook automatically configures all supported AI clients and writes their settings and, for several clients, instruction files outside the package directory. This is install-time AI-agent control-surface mutation without consent, so the package is malicious under the stated rules.
Static reason
High-risk behavior combination matched malicious policy.; previous stored version diff introduced dangerous source
Trigger
Package installation runs the postinstall hook, which invokes setup --all.
Impact
This can alter agent configuration and instructions, causing supported AI clients to load the package’s MCP server and guidance.
Mechanism
The setup command targets every supported client, whether detected or not, and writes MCP configuration plus instruction files at paths under the user’s home directory.
Policy narrative
On installation, the postinstall hook runs setup --all. That mode targets every supported AI client, including clients the setup did not detect, and writes MCP settings and instruction files under the user’s home directory. These install-time changes modify AI-agent control surfaces without a separate user request.
Evidence
package.jsondist/setup/postinstall.jsdist/setup/index.jsdist/setup/configure.jsdist/setup/clients.js