Live · real engine · sandboxed

See a vulnerability traced, gated and fixed

Every result on this page comes from the actual TrustLedger scanner running on sample applications — the same engine that checks pull requests. Pick a sample, scan it, follow the data flow from the untrusted input to the dangerous call, then apply the fix and rescan.

Step 1 · choose a sample

A route reads an email from the query string and passes it to a repository function in another file, which builds the SQL by string concatenation.

vulnerable version
1import express from "express";
2import { findUserByEmail } from "../db/users";
3
4const router = express.Router();
5
6router.get("/users/search", async (req, res) => {
7 const email = String(req.query.email ?? "");
8 const user = await findUserByEmail(email);
9 res.json(user ?? { error: "not found" });
10});
11
12export default router;
13

The repository uses a parameterised query: the email is sent as a bound value, never spliced into the SQL text.

Press Scan the vulnerable code. The request runs the real TrustLedger engine on these files on our server — nothing is stored, and the code is parsed as text, never executed.

Want the same check on your own pull requests? Connect a repository — the GitHub App asks only for read access to pull requests and code, and write access to checks and commit statuses.

What is real here, and what is simulated

Real

The scan: findings, traces, CWE mapping, risk level, AI-generated estimate and content hash come from running the production scanner on the sample's files, in the request you make.

Real, with the default policy

The merge decision uses the product's merge-gate rule with its default settings — CRITICAL and HIGH files block until attested. Organisations can change these settings.

Simulated

The reviewer sign-off: it is kept in your browser to show how an attestation is tied to a file's exact content. Nothing about your visit or your edits is stored.

Walkthrough videos

Short recordings of this demo running against the real engine. Captions are included.

TrustLedger in under a minute

A pull request with a SQL injection: the scan, the cross-file trace from the query string to the database call, and the merge held by the default policy.

Recorded from the live demo · real engine · sample code, not a customer repository

How the engine detects a real vulnerability

A multi-file flow and the step-by-step trace, then the parameterised fix side by side, the rescan, and a sanitised sample that stays clean.

Recorded from the live demo · real engine · sample code, not a customer repository

AI provenance and reviewer attestation

The AI-generated estimate and what it means, a sign-off bound to the file's content hash, and what happens to that sign-off when the code changes.

Recorded from the live demo · real engine · sample code, not a customer repository