GHSA-M547-HP4W-J6JX

Vulnerability from github – Published: 2026-03-20 14:41 – Updated: 2026-03-20 21:28
VLAI?
Summary
Vikunja has a Rate-Limit Bypass for Unauthenticated Users via Spoofed Headers
Details

Summary

Unauthenticated users are able to bypass the application's built-in rate-limits by spoofing the X-Forwarded-For or X-Real-IP headers due to the rate-limit relying on the value of (echo.Context).RealIP.

Details

In the first file below, the rate-limit for unauthenticated users can be observed being populated with the ip value. In the second file, it shows it using the c.RealIP() function for the ip case. Due to this, the rate-limit will rely on the value of one of the two mentioned headers (X-Forwarded-For or X-Real-IP). These can be spoofed by users client-side in order to completely bypass any unauthenticated rate-limits in place.

Some reverse proxies like Traefik will overwrite this value by default, but others will not, leaving any deployment that either isn't using a reserve proxy that specifically overwrites the header's value or isn't using a reverse proxy vulnerable.

File 1: pkg\routes\routes.go:318

// This is the group with no auth
    // It is its own group to be able to rate limit this based on different heuristics
    n := a.Group("")
    setupRateLimit(n, "ip")

    // Docs
    n.GET("/docs.json", apiv1.DocsJSON)
    n.GET("/docs", apiv1.RedocUI)

    // Prometheus endpoint
    setupMetrics(n)

    // Separate route for unauthenticated routes to enable rate limits for it
    ur := a.Group("")
    rate := limiter.Rate{
        Period: 60 * time.Second,
        Limit:  config.RateLimitNoAuthRoutesLimit.GetInt64(),
    }
    rateLimiter := createRateLimiter(rate)
    ur.Use(RateLimit(rateLimiter, "ip"))

File 2: pkg\routes\rate_limit.go:41

// RateLimit is the rate limit middleware
func RateLimit(rateLimiter *limiter.Limiter, rateLimitKind string) echo.MiddlewareFunc {
    return func(next echo.HandlerFunc) echo.HandlerFunc {
        return func(c *echo.Context) (err error) {
            var rateLimitKey string
            switch rateLimitKind {
            case "ip":
                rateLimitKey = c.RealIP()
            case "user":
                auth, err := auth2.GetAuthFromClaims(c)
                if err != nil {
                    log.Errorf("Error getting auth from jwt claims: %v", err)
                }
                rateLimitKey = "user_" + strconv.FormatInt(auth.GetID(), 10)
            default:
                log.Errorf("Unknown rate limit kind configured: %s", rateLimitKind)
            }

PoC

  1. Download and run the default docker compose file via the instructions here: https://vikunja.io/install/. Do not configure a proxy.
  2. Once running, navigate to the application in a web browser that is using a web proxy, such as Burp Suite.
  3. Attempt to authenticate to the application with an invalid username and password.
  4. In the web proxy's logs, locate the request to the /api/v1/login endpoint. Observe that the response contains rate-limit details:
HTTP/1.1 403 Forbidden
Cache-Control: no-store
Content-Type: application/json
Vary: Origin
X-Ratelimit-Limit: 10
X-Ratelimit-Remaining: 9
X-Ratelimit-Reset: 1772224455
Date: Fri, 27 Feb 2026 20:33:16 GMT
Content-Length: 54

{"code":1011,"message":"Wrong username or password."}
  1. Add the X-Forwarded-For header with an arbitrary value, like so: X-Forwarded-For: FakeValue. Send the request 10 times, or until the rate-limit is at zero.
  2. Modify the X-Forwarded-For headers value to be different, like so: X-Forwarded-For: NewValue.
  3. Observe that the X-Ratelimit-Remaining header's value has reset its countdown and is back at 9.

Impact

Unauthenticated users can abuse endpoints available to them for different potential impacts. The immediate concern would be brute-forcing usernames or specific accounts' passwords. This bypass allows unlimited requests against unauthenticated endpoints.

Show details on source website

{
  "affected": [
    {
      "package": {
        "ecosystem": "Go",
        "name": "code.vikunja.io/api"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0.8"
            },
            {
              "fixed": "2.2.0"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-29794"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-807"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-03-20T14:41:03Z",
    "nvd_published_at": "2026-03-20T15:16:16Z",
    "severity": "MODERATE"
  },
  "details": "### Summary\nUnauthenticated users are able to bypass the application\u0027s built-in rate-limits by spoofing the `X-Forwarded-For` or `X-Real-IP` headers due to the rate-limit relying on the value of `(echo.Context).RealIP`. \n\n### Details\nIn the first file below, the rate-limit for unauthenticated users can be observed being populated with the `ip` value. In the second file, it shows it using the `c.RealIP()` function for the `ip` case. Due to this, the rate-limit will rely on the value of one of the two mentioned headers (`X-Forwarded-For` or `X-Real-IP`). These can be spoofed by users client-side in order to completely bypass any unauthenticated rate-limits in place.\n\nSome reverse proxies like Traefik will overwrite this value by default, but others will not, leaving any deployment that either isn\u0027t using a reserve proxy that specifically overwrites the header\u0027s value or isn\u0027t using a reverse proxy vulnerable. \n\n**File 1:** pkg\\routes\\routes.go:318\n```go\n// This is the group with no auth\n\t// It is its own group to be able to rate limit this based on different heuristics\n\tn := a.Group(\"\")\n\tsetupRateLimit(n, \"ip\")\n\n\t// Docs\n\tn.GET(\"/docs.json\", apiv1.DocsJSON)\n\tn.GET(\"/docs\", apiv1.RedocUI)\n\n\t// Prometheus endpoint\n\tsetupMetrics(n)\n\n\t// Separate route for unauthenticated routes to enable rate limits for it\n\tur := a.Group(\"\")\n\trate := limiter.Rate{\n\t\tPeriod: 60 * time.Second,\n\t\tLimit:  config.RateLimitNoAuthRoutesLimit.GetInt64(),\n\t}\n\trateLimiter := createRateLimiter(rate)\n\tur.Use(RateLimit(rateLimiter, \"ip\"))\n```\n**File 2:** pkg\\routes\\rate_limit.go:41\n```go\n// RateLimit is the rate limit middleware\nfunc RateLimit(rateLimiter *limiter.Limiter, rateLimitKind string) echo.MiddlewareFunc {\n\treturn func(next echo.HandlerFunc) echo.HandlerFunc {\n\t\treturn func(c *echo.Context) (err error) {\n\t\t\tvar rateLimitKey string\n\t\t\tswitch rateLimitKind {\n\t\t\tcase \"ip\":\n\t\t\t\trateLimitKey = c.RealIP()\n\t\t\tcase \"user\":\n\t\t\t\tauth, err := auth2.GetAuthFromClaims(c)\n\t\t\t\tif err != nil {\n\t\t\t\t\tlog.Errorf(\"Error getting auth from jwt claims: %v\", err)\n\t\t\t\t}\n\t\t\t\trateLimitKey = \"user_\" + strconv.FormatInt(auth.GetID(), 10)\n\t\t\tdefault:\n\t\t\t\tlog.Errorf(\"Unknown rate limit kind configured: %s\", rateLimitKind)\n\t\t\t}\n```\n\n### PoC\n1. Download and run the default docker compose file via the instructions here: [https://vikunja.io/install/](https://vikunja.io/install/). Do not configure a proxy.\n2. Once running, navigate to the application in a web browser that is using a web proxy, such as Burp Suite.\n3. Attempt to authenticate to the application with an invalid username and password.\n4. In the web proxy\u0027s logs, locate the request to the `/api/v1/login` endpoint. Observe that the response contains rate-limit details:\n```http\nHTTP/1.1 403 Forbidden\nCache-Control: no-store\nContent-Type: application/json\nVary: Origin\nX-Ratelimit-Limit: 10\nX-Ratelimit-Remaining: 9\nX-Ratelimit-Reset: 1772224455\nDate: Fri, 27 Feb 2026 20:33:16 GMT\nContent-Length: 54\n\n{\"code\":1011,\"message\":\"Wrong username or password.\"}\n```\n5. Add the `X-Forwarded-For` header with an arbitrary value, like so: `X-Forwarded-For: FakeValue`. Send the request 10 times, or until the rate-limit is at zero.\n6. Modify the `X-Forwarded-For` headers value to be different, like so: `X-Forwarded-For: NewValue`.\n7. Observe that the `X-Ratelimit-Remaining` header\u0027s value has reset its countdown and is back at `9`. \n\n### Impact\nUnauthenticated users can abuse endpoints available to them for different potential impacts. The immediate concern would be brute-forcing usernames or specific accounts\u0027 passwords. This bypass allows unlimited requests against unauthenticated endpoints.",
  "id": "GHSA-m547-hp4w-j6jx",
  "modified": "2026-03-20T21:28:21Z",
  "published": "2026-03-20T14:41:03Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/go-vikunja/vikunja/security/advisories/GHSA-m547-hp4w-j6jx"
    },
    {
      "type": "ADVISORY",
      "url": "https://nvd.nist.gov/vuln/detail/CVE-2026-29794"
    },
    {
      "type": "WEB",
      "url": "https://github.com/go-vikunja/vikunja/commit/a498dd69915a006c07e9d82660a2185d7e8136ee"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/go-vikunja/vikunja"
    },
    {
      "type": "WEB",
      "url": "https://vikunja.io/changelog/vikunja-v2.2.0-was-released"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:L",
      "type": "CVSS_V3"
    }
  ],
  "summary": "Vikunja has a Rate-Limit Bypass for Unauthenticated Users via Spoofed Headers"
}


Log in or create an account to share your comment.




Tags
Taxonomy of the tags.


Loading…

Loading…

Loading…

Sightings

Author Source Type Date

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…

Detection rules are retrieved from Rulezet.

Loading…

Loading…