GHSA-JFGX-762W-99VR

Vulnerability from github – Published: 2026-10-20 20:37 – Updated: 2026-10-20 20:37
VLAI
Summary
@angular/common: XSRF Token Leakage via URL Parsing Discrepancy in Angular HTTP Client (Incomplete fix for CVE-2025-66035)
Details

An incomplete fix for GHSA-58c5-g7wp-6w37 (CVE-2025-66035) in @angular/common's HTTP Client (HttpClient) allows Cross-Site Request Forgery (XSRF) tokens to be leaked to attacker-controlled cross-origin domains when requests use URLs that exploit parsing discrepancies between regular expressions and the browser's WHATWG URL parser.

Technical Description

To address CVE-2025-66035, xsrfInterceptorFn introduced a regular expression (const ABSOLUTE_URL_REGEX = /^(?:https?:)?\/\//i) to detect and skip absolute and protocol-relative URLs before attaching the X-XSRF-TOKEN header.

However, browsers resolve request URLs passed to fetch() and XMLHttpRequest.open() using the WHATWG URL parser. For http: and https: schemes, the WHATWG URL standard performs several normalizations prior to resolving the authority/origin: * Backslashes (\) are treated as forward slashes (/) — e.g., /\evil.example/x, \\evil.example/x, \/evil.example/x, https:\\evil.example/x, https:/\evil.example/x. * Leading and trailing C0 control characters and spaces (<= 0x20) are stripped — e.g., ' //evil.example/x', '\0//evil.example/x', '\x1f//evil.example/x'. * ASCII tab (\t), newline (\n), and carriage return (\r) characters are removed anywhere in the input — e.g., '/\t/evil.example/x', '/\n/evil.example/x', 'htt\tps://evil.example/x'.

Because these URLs do not match /^(?:https?:)?\/\//i, xsrfInterceptorFn treats them as relative same-origin URLs and attaches the X-XSRF-TOKEN header, while the browser resolves and dispatches the request to the cross-origin host (evil.example).

Note: Angular 21.0.3+ and 22.x replaced the regular expression with a WHATWG new URL(req.url, locationOrigin) origin comparison in #65683 and were already unaffected. However, the 20.3.x LTS branch retained the regular expression until 20.3.34.

Impact

The token leakage bypasses Angular's built-in XSRF protection on affected versions, allowing an attacker to capture the user's valid X-XSRF-TOKEN header if their server permits the request and custom header via CORS. Once obtained, the attacker can perform Cross-Site Request Forgery (CSRF) attacks against the victim user's session.

Attack Preconditions

  1. The victim's Angular application must have XSRF protection enabled (enabled by default in HttpClient).
  2. The attacker must be able to make the application send a state-changing HTTP request (e.g., POST, PUT, DELETE, PATCH) to a URL they control (such as /\attacker.example/api or //attacker.example/api).

Patches

The issue is resolved by parsing and comparing the request URL's origin against the application's current origin using the standard URL constructor (new URL(req.url, locationOrigin)).

  • 21.0.3 (and all 22.x releases)
  • 20.3.34

Workarounds

If you cannot immediately upgrade to a patched version: * Validate and normalize outgoing request URLs using new URL(url, window.location.origin) and verify that parsedUrl.origin === window.location.origin before passing dynamic URLs to mutating HttpClient requests. * Avoid passing untrusted or user-controlled URLs to HttpClient mutating methods.

Severity
5.9 (Medium)
CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:A/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N
Show details on source website
To clipboard 

{
  "affected": [
    {
      "package": {
        "ecosystem": "npm",
        "name": "@angular/common"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "21.0.0"
            },
            {
              "fixed": "21.0.3"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    },
    {
      "package": {
        "ecosystem": "npm",
        "name": "@angular/common"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "20.0.0"
            },
            {
              "fixed": "20.3.34"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    },
    {
      "package": {
        "ecosystem": "npm",
        "name": "@angular/common"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "last_affected": "19.2.25"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-111753"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-201",
      "CWE-359"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-10-20T20:37:57Z",
    "nvd_published_at": null,
    "severity": "MODERATE"
  },
  "details": "An incomplete fix for [GHSA-58c5-g7wp-6w37](https://github.com/angular/angular/security/advisories/GHSA-58c5-g7wp-6w37) (`CVE-2025-66035`) in `@angular/common`\u0027s HTTP Client (`HttpClient`) allows Cross-Site Request Forgery (XSRF) tokens to be leaked to attacker-controlled cross-origin domains when requests use URLs that exploit parsing discrepancies between regular expressions and the browser\u0027s WHATWG URL parser.\n\n### Technical Description\n\nTo address `CVE-2025-66035`, `xsrfInterceptorFn` introduced a regular expression (`const ABSOLUTE_URL_REGEX = /^(?:https?:)?\\/\\//i`) to detect and skip absolute and protocol-relative URLs before attaching the `X-XSRF-TOKEN` header.\n\nHowever, browsers resolve request URLs passed to `fetch()` and `XMLHttpRequest.open()` using the **WHATWG URL parser**. For `http:` and `https:` schemes, the WHATWG URL standard performs several normalizations prior to resolving the authority/origin:\n* Backslashes (`\\`) are treated as forward slashes (`/`) \u2014 e.g., `/\\evil.example/x`, `\\\\evil.example/x`, `\\/evil.example/x`, `https:\\\\evil.example/x`, `https:/\\evil.example/x`.\n* Leading and trailing C0 control characters and spaces (`\u003c= 0x20`) are stripped \u2014 e.g., `\u0027 //evil.example/x\u0027`, `\u0027\\0//evil.example/x\u0027`, `\u0027\\x1f//evil.example/x\u0027`.\n* ASCII tab (`\\t`), newline (`\\n`), and carriage return (`\\r`) characters are removed anywhere in the input \u2014 e.g., `\u0027/\\t/evil.example/x\u0027`, `\u0027/\\n/evil.example/x\u0027`, `\u0027htt\\tps://evil.example/x\u0027`.\n\nBecause these URLs do not match `/^(?:https?:)?\\/\\//i`, `xsrfInterceptorFn` treats them as relative same-origin URLs and attaches the `X-XSRF-TOKEN` header, while the browser resolves and dispatches the request to the cross-origin host (`evil.example`).\n\nNote: Angular `21.0.3+` and `22.x` replaced the regular expression with a WHATWG `new URL(req.url, locationOrigin)` origin comparison in [#65683](https://github.com/angular/angular/pull/65683) and were already unaffected. However, the `20.3.x` LTS branch retained the regular expression until `20.3.34`.\n\n### Impact\n\nThe token leakage bypasses Angular\u0027s built-in XSRF protection on affected versions, allowing an attacker to capture the user\u0027s valid `X-XSRF-TOKEN` header if their server permits the request and custom header via CORS. Once obtained, the attacker can perform Cross-Site Request Forgery (CSRF) attacks against the victim user\u0027s session.\n\n### Attack Preconditions\n\n1. The victim\u0027s Angular application must have **XSRF protection enabled** (enabled by default in `HttpClient`).\n2. The attacker must be able to make the application send a state-changing HTTP request (e.g., `POST`, `PUT`, `DELETE`, `PATCH`) to a URL they control (such as `/\\attacker.example/api` or ` //attacker.example/api`).\n\n### Patches\n\nThe issue is resolved by parsing and comparing the request URL\u0027s origin against the application\u0027s current origin using the standard `URL` constructor (`new URL(req.url, locationOrigin)`).\n\n* `21.0.3` (and all `22.x` releases)\n* `20.3.34`\n\n### Workarounds\n\nIf you cannot immediately upgrade to a patched version:\n* Validate and normalize outgoing request URLs using `new URL(url, window.location.origin)` and verify that `parsedUrl.origin === window.location.origin` before passing dynamic URLs to mutating `HttpClient` requests.\n* Avoid passing untrusted or user-controlled URLs to `HttpClient` mutating methods.",
  "id": "GHSA-jfgx-762w-99vr",
  "modified": "2026-10-20T20:37:57Z",
  "published": "2026-10-20T20:37:57Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/angular/angular/security/advisories/GHSA-jfgx-762w-99vr"
    },
    {
      "type": "WEB",
      "url": "https://github.com/angular/angular/pull/65683"
    },
    {
      "type": "WEB",
      "url": "https://github.com/angular/angular/pull/71295"
    },
    {
      "type": "WEB",
      "url": "https://github.com/angular/angular/commit/0659d11c854c9c6b886a2e925afef98d70bb005f"
    },
    {
      "type": "WEB",
      "url": "https://github.com/angular/angular/commit/20474d3f0fd7c64071add6e84acf720627e5c19b"
    },
    {
      "type": "WEB",
      "url": "https://github.com/angular/angular/commit/933ce817d53e25926844fa56af8230816763d08e"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/angular/angular"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:A/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N",
      "type": "CVSS_V4"
    }
  ],
  "summary": "@angular/common: XSRF Token Leakage via URL Parsing Discrepancy in Angular HTTP Client (Incomplete fix for CVE-2025-66035)"
}



Log in or create an account to share your comment.




Tags
Taxonomy of the tags.




Drag nodes to arrange them. Double-click a related vulnerability to open it, or a source to visit its link. The Physics button on the left tunes the layout and the simulation; the legend at the bottom left hides or shows nodes by kind and by sighting type.

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.
Vulnerability timeline

Sightings and recorded dates, grouped by day · UTC

JSON
How durations are calculated

Following the VTEM framework (Black Hat USA 2015), time to patch starts when the vendor or a coordinator was informed, total time to patch at discovery, and disclosure to patch at public disclosure; vendor response time runs from the report to its acknowledgement. Time to exploit and time unpatched start at the earliest discovery or report, falling back to public disclosure. Unpatched time ends at the first recorded patch, or the calculation time when no patch is recorded. Exploitation to patch shows the interval between those two events; a published proof of concept is a public exploit, not exploitation.

Each result shows its source dates. A negative interval means the end event came first; when date-only inputs leave the order open, the result says so. Date-only inputs are approximate. The earliest evidence wins: KEV report dates and sightings are labelled as such. Records describing another vulnerability are listed but never calculated from, and a PoC alone does not establish exploitation. Missing patch evidence does not prove that no patch exists. Filters below change the event list only.

Loading timeline…




    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.