AI Rescue · Free Auto Scan
Built it with Cursor or Claude? Free scan, A–F in 30 minutes.
Submit a GitHub URL. Get a code-health grade, the top critical issues, and a recommended fix tier — by email. No card required, no sales call attached.
Common situations
"We built it with AI and now we are not sure it is safe to keep running."
AI Rescue is for the situations the Vibe Coding wave created — code that works but nobody can confidently say is OK.
Built an app with Cursor or Claude. It works. You cannot tell if it is safe in production.
Security audits start at $5K. This is your $0 first read.It worked last week, then stopped. You do not know what broke, or where to start.
An A–F report with the top three Critical issues turns "I don't know" into a task list.A non-technical co-founder needs to explain the codebase to investors. There is no document.
The Diagnostic ($500) is a written technical report — designed to share, not to be deciphered.Before → After
What changes after 30 minutes.
The Auto Scan does not fix anything. It tells you exactly what to worry about, with priority. That is the change.
- "It works, but is it really OK?" — and no one in-house can answer.
- It stopped working and nobody knows what broke.
- Security worries, no engineer to ask.
- "Probably fine" is the best you can say to a board, an investor, or a customer.
- A–F grade for the codebase, in writing.
- Critical / High / Medium issues counted and named — file and line.
- Top concern called out with sample fix code.
- "What to do next" — DIY guide, paid diagnostic, or staged fix.
Sample report
What your Auto Scan email actually looks like.
Same structure, every time. A grade, a count, a top concern with a file and line, and a "what next" path.
/api/users.js (line 42) — direct string interpolation in query.Sample fix code included in this email.config/email.js (line 7) — committed to public history.Rotate the key before any other fix./api/upload — accepts arbitrary file types.Pre-built validator linked in the DIY guide.5 staged tiers
Free at the entry. Pay only for what you decide to fix.
The ladder is designed to be exited at any rung. Most clients stop at the diagnostic; some never need more than the free scan.
Auto Scan
Automated AI scan + static analysis. A–F grade, top critical issues by file and line. Email-delivered.
Diagnostic Report
Detailed human + AI analysis. Technical report, sample fix code, fix plan with priority and estimates. NDA-signed.
Hotfix
Critical issues only. Minimum viable working state. Team: PM + Engineer + QA.
Stabilize
Hotfix + High fixes + tests (50%+ coverage) + CI/CD + basic monitoring. Stable production state.
Rebuild
Full architecture redesign, migration plan, phased rollout. For codebases beyond stabilising.
How the Auto Scan works
What happens after you submit the form.
The Auto Scan is automated end-to-end. No human reads your code until you choose the Diagnostic tier — and only then under NDA.
You submit the form
5 fields: GitHub URL, what you built it with (Cursor / Claude / GPT / Copilot / Other), current state, production yes/no, email. Takes a minute.
No card required
Repo cloned (read-only)
We clone your repo read-only. Static analysis tools run in parallel — language linters, dependency audit, secret detection, security pattern scan.
ESLint · npm audit · gitleaks · Semgrep
AI analysis layer
Claude analyses the static-analysis output plus the code itself, prioritises issues, drafts sample fix code, generates the email content. No human reads the code at this stage.
Fully automated
Email delivered, repo deleted
The email arrives in your inbox. The cloned repo is deleted from our infrastructure. Your email address is on file for follow-up unless you opt out.
Repo deleted · Email on file
Why this is free
The numbers behind $0.
Auto Scan is free because 80% of diagnosis is AI-automated, and because most scans do not need a human to read the code at all. We run the scans at our cost.
Security audits start at $5,000. The Auto Scan is $0. The Diagnostic ($500) is one-tenth the entry price of a typical audit, with output that includes fix code samples.
Audits take 1–4 weeks. Auto Scan takes 10–30 minutes. The speed comes from AI doing the routine work; humans are reserved for the Diagnostic and beyond.
Tsunagu, Dr. Unco, and L-career — live codebases we build and operate ourselves. We know what AI-generated code looks like in production because we run it every day.
What you can rely on
Five rules for every Auto Scan.
Operational rules — applied on every scan, free or paid.
Read-only clone
We clone your repo with read-only access. We cannot modify your code, your repository settings, or your secrets.
Repo deleted post-scan
The cloned repository is deleted from our infrastructure once the scan email is sent. No code retained.
No human reads code (free tier)
The Auto Scan is end-to-end automated. A human only reads your code if you choose the Diagnostic tier, and only after NDA is signed.
No sales chase
Day 3, 2-week, and 3-month follow-ups are sent only if you do not reply. Unsubscribe is one click. We do not call your phone.
NDA before tech (paid tiers)
Every paid tier — Diagnostic, Hotfix, Stabilize, Rebuild — starts with NDA and DPA before any technical discussion or human code review.
FAQ
Five questions before you scan.
Specific case not here? Add it to the form's message field. Reply within one business day.
Is the Auto Scan really free?
What if the scan shows everything is fine?
How is this different from a security audit?
What does each fix tier do?
What about the code I built with Cursor / Claude?
Free Auto Scan
Paste a GitHub URL. Get a code-health grade by email.
No card. No call. 10–30 minutes from submit to result.
- Read-only clone, deleted after the scan completes.
- End-to-end automated — no human reads your code at the free tier.
- Email follow-ups only if you do not reply; unsubscribe in one click.
- $0
- 10–30 min
- Repo deleted
- No card
Scan my code — free
Required fields marked * · 5 fields, 30 seconds