The package configures project-level agent hooks and instructions through its CLI, so it presents a first-party agent extension setup risk. The manifest has no npm install lifecycle trigger, and the inspected code did not establish malicious behavior.
Static reason
High-risk behavior combination matched malicious policy.; previous stored version diff introduced dangerous source
Trigger
A user runs the CLI and selects project setup or launch with updates, or runs migrate --apply.
Impact
The configured agent tools can run the bundled hooks and load the package's instructions in later sessions; no credential theft or unconsented npm install-time mutation was identified.
Mechanism
The package writes Claude settings and hooks and exports instructions or skills into project-level agent configuration locations.
Evidence
package.jsondist/index.jsdist/setup.jsdist/mcp.jsdist/migrate.jsdist/targets/project-ui.jsdist/targets/agents-md.jsdist/targets/cursor-rules.jsdist/targets/grok-agents.jsdist/targets/grok-skills.js