Find AI-generated production gaps that tests miss, then keep a verification trail of what was found and fixed. No GitHub OAuth. Source stays on your machine.
PAYMENTS_URL = os.getenv(
"PAYMENTS_URL",
"https://mock-payments-stage.example"
)Tests pass. The application works.
But production can silently use a test payment system.
CRITICAL — Production fallback points to staging payment infrastructure.
Tests tell you whether code works. SimpleBeacon checks whether it works against the reality you intended.
Try it locally. No account. Paid CI gates come later, after you have seen a finding on your own repo.
Your AI coding agent can make your tests pass—and still wire production to the wrong reality. Existing tools tell you whether the code works. SimpleBeacon checks the assumptions underneath it.
const METRICS_DATA_PATH =
'./fixtures/sample-metrics.json';
const conversionRate = '75%';Most developers: “It’s demo code.”
This code is in a production handler.
CRITICAL — Production sample-data dependency
CRITICAL — Hard-coded KPI
That’s what SimpleBeacon is for.
The dangerous generated code isn’t ugly. It’s believable code that isn’t true.
Staging or mock infrastructure wired into production fallbacks and defaults.
Fixtures or fabricated metrics presented as live data in production handlers.
Placeholders and no-op implementations that look finished and ship anyway.
APIs and packages that don’t actually exist, with completely reasonable names.
SimpleBeacon is zero-custody: source code, file bytes, and full finding lists stay on the machine that ran the scan. The cloud only sees account details you submit (email, product, scan list) and an optional hex scan fingerprint of counts plus project name—not the repo and not the finding list. Engineers can evaluate the CLI locally, then hand a verification report or this boundary to a manager. It is a verification trail, not a legal conformity certificate.
report.jsonscanHash (hex, at least 16 characters)Inspect the CLI on GitHub. Paste a local JSON report on Roadmap without uploading it. Scoped CLI gate: shipping-path scan.
If you need evidence of how you control AI-generated software, SimpleBeacon produces an audit trail of what it found, what was fixed, and when it was fixed. That’s a verification report—not a legal conformity certificate.
Scan locally, fix the finding, paste a gate badge on the pull request. No OAuth to try it.
After a local scan, buy an Executive Risk Certificate pack for a risk committee, or start an EU AI Act sprint. Same engine. Paid step is the signed pack and CI enforcement—not the first scan.
$499 Executive Risk Certificate
$2,499 EU AI Act Readiness Sprint
$50,000+ annual enterprise contracts
Scan locally. No upload required. Automating the same gate on pull requests is the paid team step.