Source inspection shows explicit-command setup of AI hooks that can capture and upload prompt and response content, with a notice before configuration changes. This supports a warning for agent capability risk, not a finding of an unconsented npm install attack.
Static reason
No blocking static signals were detected.; previous stored version diff introduced dangerous source
Trigger
The user runs `litmus init` with an assessment token.
Impact
AI prompt and response content can be recorded and sent to the assessment service; this is a disclosed, package-owned setup but is a sensitive agent capability.
Mechanism
The command installs hooks for multiple AI tools and marks some assessment folders trusted, enabling capture without a separate per-hook approval when trust setup succeeds.
Evidence
dist/commands/init.jsdist/lib/ai-tracking.jsdist/lib/hook-logger.cjs