These samples preserve output originally captured with CLI 1.5.3 on the example template in examples/nextjs-supabase-vercel-production-ready-template, which holds agent rule files and no app code. The counts and findings are historical; they have not been recaptured with the current release.
The local project path is replaced with .. The public repo export adapts reproduction and follow-up npx commands in these samples to the current release; the CLI originally printed them without a version. The command header in terminal-scan.sample.txt is a current reproduction instruction, followed by the original CLI 1.5.3 stdout with its local path replaced. In agent-tasklist.sample.md, the dash the CLI prints in “No automated recipe; see scanner hint.” is written as a semicolon. No secrets or private project IDs.
To collect a fresh result, run this in the example template folder (local, no login), then compare .viberaven/gate-result.json and .viberaven/agent-tasklist.md with the historical samples. A current release can return different findings and counts:
npx -y viberaven@1.6.7 scan
Files:
gate-result.sample.json: original machine verdict (gate.status is not_clear)agent-tasklist.sample.md: original task list with exported follow-up commands adapted to the current releaseterminal-scan.sample.txt: current reproduction command header and original CLI 1.5.3 stdoutThe scan reads repo files only. On a template with no app code, some stack rows name tools the template does not use, such as Vue and Netlify. A clear result covers the repo checks only; it is not a security audit or a live check of Supabase or Vercel.
Find missing RLS before deployment compares two small synthetic fixtures using the published CLI. It includes the SQL change, expected findings and exit codes, and the live-database verification boundary.
Review an AI-built app before launch or client handoff connects repository findings to a blank evidence record, with prompts for a readiness check, a Lovable + Supabase launch, and a client handoff.