A $1 AI-checked security scan, sold as a launch-day review
Security scanners bury small teams in false alarms. A new scanner has Claude judge every finding and write the fix, for about a dollar a scan, which leaves the valuable part, confirming what's real, as a service.
Published Oct 24, 2026
The Wow
What someone built
OpenRouting runs Semgrep's open rules on a GitHub repo in a sealed container, then routes each finding to a specialist model reviewer for a verdict, a confidence score, a plain-English explanation and a fix. It also checks AI-agent setup files and does a passive check of the live site.
- License
- Hosted commercial service paid in credits (100 free at sign-up), not open source. We link to it; we don't copy it
- To use it
- Nothing to install: sign in at openrouting.si with Google or GitHub, connect a repository and start a scan.
- A business you could build on it
- Launch-day security review for AI-built apps
Why it's remarkable
By list price ($1 for 20 credits, 20 per full scan) the scan costs about a dollar, and it reviews the instruction files and MCP configs that steer coding agents, which most scanners skip. The creator lists what it doesn't do yet instead of hiding it.
The Opportunity
The business it reveals
A launch-day security review for founders shipping AI-built apps: scan, confirm by hand, explain the three issues that matter and re-check the fixes. The tool does the sorting; you sell the judgment.
Who would pay
Solo founders and small teams launching apps built with AI coding agents, and the agencies that build apps for clients.
The Flip
3 ways you could turn this into a business
Launch-day security review for AI-built apps
- Who pays
- Solo founders and small teams launching apps built with AI coding agents
- How it would earn
- A fixed fee per review plus a paid re-check after fixes (our hypothesis; price untested)
- Why it might work
- The scan costs about a dollar by list price, so nearly all the value is the confirmation and the explanation, which is exactly what a founder without a security person lacks.
- What already exists
- Snyk, Semgrep and GitHub Advanced Security sell scanners to teams; freelance penetration testers sell deep tests at much higher prices. A short confirmed review sold to non-specialists sits between them.
Security handover pack for app-building agencies
- Who pays
- Freelancers and small agencies that build apps for clients
- How it would earn
- A per-project add-on the agency resells to its client (our hypothesis)
- Why it might work
- Agencies need to show clients the code was checked; a dated report and a fix log is a document they can attach to the handover.
- What already exists
- Agencies running scanners themselves or skipping the check; few small shops have a reviewer on staff.
AI-agent setup audit for teams using coding agents
- Who pays
- Engineering leads at small companies rolling out Claude Code, Codex or Cursor
- How it would earn
- A one-time audit of agent instructions, hooks and MCP connections (our hypothesis)
- Why it might work
- The tool already checks CLAUDE.md, skills, hooks and MCP configs for prompt-injection and tool-safety risks, a newer surface most teams haven't reviewed at all.
- What already exists
- Internal security teams at larger companies; we found few services aimed at small teams for this, so treat demand as untested.
Why we'd try "Launch-day security review for AI-built apps" first: The buyer has a deadline and a clear fear, the scan is close to free, and your work is the part the tool can't own: confirming what's real and explaining it.
Build It
How you could build it
A fixed-scope launch review: repo scan, passive site scan and agent-setup check, hand-confirmed findings, a dependency check, a two-page report and one re-check.
First step: Scan one of your own repositories with the free credits, then check every finding it calls real or false by hand. That tells you how much of a review the tool does and how much you would be selling.
Agents and tools
- OpenRouting
- GitHub
- npm audit, pip-audit or Dependabot
- An AI agent to adapt prompts and check the work (ChatGPT, Claude, Gemini, Grok Bot or similar)
Setup
- Run the scanner on two or three of your own or open-source repos and grade its verdicts.
- Write a permission form and a report template that lists what was and wasn't checked.
- Add a dependency check, since the scanner doesn't do one yet.
- Set one fixed price and a re-check price.
What the agent does
- Runs the rules and reviews each finding
- Drafts explanations and fixes
- Drafts the report summary
What you do
- Gets written permission
- Confirms every finding
- Runs the dependency check
- Explains the fixes to the client and re-checks them
Your first customer: A founder in your network who is launching soon, offered a free first review for a testimonial.
Test demand first: Review three repos, including one of your own with a planted bug. Success: the planted issue is caught, every confirmed finding holds up, and at least one founder asks what the review would cost.
Your starter instructions
Paste this into the AI agent you use (ChatGPT, Claude, Gemini, Grok Bot or another) as your first message. Copying it starts work in your own agent; it doesn't create or deploy a business by itself.
Running it
What stays human: Getting written permission to scan, confirming every finding the model marks as real, adding dependency checks the tool doesn't do yet, writing the summary in the client's terms, and walking the founder through the fixes.
- Startup cost (estimate)
- About $0 to start (estimate) (Estimate from the creator's pricing page: sign-up includes 100 free credits, enough for five full scans. Not measured)
- Running cost
- Roughly $1 to $3 in credits per client (estimate) (Rough arithmetic from the list price of $1 per 20 credits: a full scan is 20 credits, a PDF report 10, a passive site scan 10 and an agent-setup review 10 more. Your hours confirming findings are the real cost and are not measured)
- How much the agent runs
- People run it; AI helps with tasks.
Pricing your own version? Use the free startup-cost worksheet.
The evidence
What the source shows: The creator's DEV post and pricing page (Oct 24, 2026). We read both; we haven't run a scan.
What's ours: The review service, its pricing and the flips are our application.
Worth planning around
- The product launched publicly on Oct 24, 2026 with no outside reviews yet.
- No dependency (CVE) or git-history scanning yet, per the creator.
- Code snippets go to a model on Amazon Bedrock; clients should know before you connect a repo.
Sources
- OpenRouting: I built a security scanner where an AI reviews every finding (DEV, Oct 24, 2026)
- OpenRouting on DEV
- OpenRouting pricing
Sources reopened Oct 24, 2026. Spot a mistake? Tell us and we'll correct it in public.
Keep reading
- Services to sell to people building apps with AI (guide)
- Cursor: plans and what people build and sell with it (guide)
- How much of the work do AI agents actually run? (guide)
- He still can't write code. His five-person ad agency now runs on about 270 tools he built with Claude. Every owner-run firm needs its first five
- A marketer who doesn't code built an elk-call coaching app with Claude Code and says it hit a $650K run rate in 110 days. Every skill learned by sound is the same opening
- A free n8n workflow flags fake supplier invoices before anyone pays them — and Germany's e-invoice switch makes it a service bookkeepers can sell
- Marc Lou's agent chased a broken-shoe refund through three support desks and got his 850 AED back. Small stores now need a claims desk that can answer agents
- Every discovery and launch plan