The package performs purpose-related, disclosed capture through an explicit assessment setup command, so the source does not establish malicious behavior. However, setup changes global AI-agent configuration and enables capture outside the assessment folder, making this a meaningful capability users should be warned about.
Static reason
One or more suspicious static signals were detected.; previous stored version diff introduced dangerous source
Trigger
The user runs `litmus init` for an assessment whose capture setup is deferred; the source indicates the disclosure was already accepted on the assessment page.
Impact
The assessment company can receive AI prompts, responses, and tool activity from supported tools used outside the assessment folder during the capture session.
Mechanism
The CLI writes hooks into user-level AI tool configuration, including a Codex trust pin, then the hook logger posts captured prompt and response text to `/cli/activity` with the assessment token.
Evidence
dist/commands/init.jsdist/lib/hosted-capture.jsdist/lib/ai-tracking.jsdist/lib/hook-logger.cjs~/.litmus/hosted-capture.json~/.litmus/hook-logger.cjs~/.claude/settings.json~/.codex/hooks.json