Vulnerability-Lookup

GHSA-29RF-F4VV-PVQ6

Vulnerability from github – Published: 2026-08-14 15:36 – Updated: 2026-08-14 15:36

Summary

Authorizer: Zero-click account takeover via OAuth identity linking to unverified email accounts

Details

The OAuth callback handler links incoming OAuth identities (Google, GitHub, etc.) to existing accounts matched by email address without verifying that the existing account's email was verified by its original owner. An attacker who pre-registers with a victim's email address (without verifying it) gains persistent password-based access to the victim's account after the victim completes a normal OAuth login. Verified against HEAD (commit 73679fa).

Root Cause

In internal/http_handlers/oauth_callback.go, when an OAuth login occurs for an email that already exists in the database:

Line 125: The existing user is looked up by email:

existingUser, err := h.StorageProvider.GetUserByEmail(ctx, refs.StringValue(user.Email))

Line 164: The OAuth user object is replaced with the existing user:

user = existingUser

Lines 173-176: The OAuth provider is appended to the existing user's signup methods:

signupMethod := existingUser.SignupMethods
if !strings.Contains(signupMethod, provider) {
    signupMethod = signupMethod + "," + provider
}
user.SignupMethods = signupMethod

Lines 179-181: If the existing account's email was NOT verified, it is automatically verified:

if user.EmailVerifiedAt == nil {
    now := time.Now().Unix()
    user.EmailVerifiedAt = &now
}

Line 219: The merged user is saved to the database:

user, err = h.StorageProvider.UpdateUser(ctx, user)

At no point is the existing account's password invalidated or the owner notified that a new OAuth identity was linked.

Attack Chain

Attacker signs up with victim@company.com using email/password. Attacker sets a known password but does NOT click the email verification link. The account exists in the database with EmailVerifiedAt = nil.
Some time later, the real owner of victim@company.com logs in via Google OAuth (a completely normal action).
The OAuth callback at line 125 finds the attacker's existing account by email.
At line 164, the Google OAuth identity is linked to the attacker's account.
At line 179-181, the email is automatically verified (the attacker never verified it, but now it's marked as verified).
At line 175, "google" is appended to the signup methods. The account now has both "basic_auth" and "google" as valid login methods.
The attacker's original password is still valid in the database. It was never cleared, changed, or invalidated.
The attacker logs in with victim@company.com and the password they originally set. They now have full access to the victim's account, including any data the victim added via their Google session.

Why This Is Zero-Click

The victim performs no unusual action. They simply log in via their Google account, which is the expected, secure behavior. The attacker staged the account beforehand and gains access without any further interaction.

This is a classic Account Linking vulnerability (cited in OWASP authentication guidelines). The core logic flaw is a trust boundary violation. Authorizer correctly trusts that Google has verified the email address, but it incorrectly extends that trust to validate the password that was set by the unverified attacker. The attacker maintains persistent, password-based backdoor access to the victim's account, even if the victim later revokes Authorizer's OAuth access from their Google account settings. The password was set before Google was ever involved and is never invalidated by the linking process.

Impact

Full account takeover for any user who logs in via OAuth
Attacker maintains persistent password-based access even after the victim changes OAuth providers
All data the victim creates after OAuth login is accessible to the attacker
The victim has no indication their account was pre-staged
Affects every OAuth provider configured in Authorizer (Google, GitHub, Facebook, Apple, LinkedIn, Twitter, Discord, Twitch, Roblox, Microsoft)

Suggested Fix

Before linking an OAuth identity to an existing account, verify that the existing account's email is already verified:

existingUser, err := h.StorageProvider.GetUserByEmail(ctx, refs.StringValue(user.Email))
if err == nil {
    // Account exists. Only link if email is already verified.
    if existingUser.EmailVerifiedAt == nil {
        // Email not verified by original owner. Do NOT link.
        // Either: reject the login, or create a new separate account,
        // or delete the unverified account and create a fresh one for the OAuth user.
    }
}

Additionally, when linking a new OAuth identity, invalidate any existing password on the account or require the user to re-authenticate via the original method.

Credit

Koda Reef

Severity

8.7 (High)


                  
                    CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N

Show details on source website

JSON

To clipboard

{
  "affected": [
    {
      "package": {
        "ecosystem": "Go",
        "name": "github.com/authorizerdev/authorizer"
      },
      "ranges": [
        {
          "events": [
            {
              "introduced": "0"
            },
            {
              "fixed": "0.0.0-20260807033110-66fe488fd2a4"
            }
          ],
          "type": "ECOSYSTEM"
        }
      ]
    }
  ],
  "aliases": [
    "CVE-2026-35511"
  ],
  "database_specific": {
    "cwe_ids": [
      "CWE-287"
    ],
    "github_reviewed": true,
    "github_reviewed_at": "2026-08-14T15:36:15Z",
    "nvd_published_at": null,
    "severity": "HIGH"
  },
  "details": "The OAuth callback handler links incoming OAuth identities (Google, GitHub, etc.) to existing accounts matched by email address without verifying that the existing account\u0027s email was verified by its original owner. An attacker who pre-registers with a victim\u0027s email address (without verifying it) gains persistent password-based access to the victim\u0027s account after the victim completes a normal OAuth login. Verified against HEAD (commit 73679fa).\n\n## Root Cause\n\nIn `internal/http_handlers/oauth_callback.go`, when an OAuth login occurs for an email that already exists in the database:\n\nLine 125: The existing user is looked up by email:\n\n    existingUser, err := h.StorageProvider.GetUserByEmail(ctx, refs.StringValue(user.Email))\n\nLine 164: The OAuth user object is replaced with the existing user:\n\n    user = existingUser\n\nLines 173-176: The OAuth provider is appended to the existing user\u0027s signup methods:\n\n    signupMethod := existingUser.SignupMethods\n    if !strings.Contains(signupMethod, provider) {\n        signupMethod = signupMethod + \",\" + provider\n    }\n    user.SignupMethods = signupMethod\n\nLines 179-181: If the existing account\u0027s email was NOT verified, it is automatically verified:\n\n    if user.EmailVerifiedAt == nil {\n        now := time.Now().Unix()\n        user.EmailVerifiedAt = \u0026now\n    }\n\nLine 219: The merged user is saved to the database:\n\n    user, err = h.StorageProvider.UpdateUser(ctx, user)\n\nAt no point is the existing account\u0027s password invalidated or the owner notified that a new OAuth identity was linked.\n\n## Attack Chain\n\n1. Attacker signs up with `victim@company.com` using email/password. Attacker sets a known password but does NOT click the email verification link. The account exists in the database with `EmailVerifiedAt = nil`.\n\n2. Some time later, the real owner of `victim@company.com` logs in via Google OAuth (a completely normal action).\n\n3. The OAuth callback at line 125 finds the attacker\u0027s existing account by email.\n\n4. At line 164, the Google OAuth identity is linked to the attacker\u0027s account.\n\n5. At line 179-181, the email is automatically verified (the attacker never verified it, but now it\u0027s marked as verified).\n\n6. At line 175, \"google\" is appended to the signup methods. The account now has both \"basic_auth\" and \"google\" as valid login methods.\n\n7. The attacker\u0027s original password is still valid in the database. It was never cleared, changed, or invalidated.\n\n8. The attacker logs in with `victim@company.com` and the password they originally set. They now have full access to the victim\u0027s account, including any data the victim added via their Google session.\n\n## Why This Is Zero-Click\n\nThe victim performs no unusual action. They simply log in via their Google account, which is the expected, secure behavior. The attacker staged the account beforehand and gains access without any further interaction.\n\nThis is a classic Account Linking vulnerability (cited in OWASP authentication guidelines). The core logic flaw is a trust boundary violation. Authorizer correctly trusts that Google has verified the email address, but it incorrectly extends that trust to validate the password that was set by the unverified attacker. The attacker maintains persistent, password-based backdoor access to the victim\u0027s account, even if the victim later revokes Authorizer\u0027s OAuth access from their Google account settings. The password was set before Google was ever involved and is never invalidated by the linking process.\n\n## Impact\n\n- Full account takeover for any user who logs in via OAuth\n- Attacker maintains persistent password-based access even after the victim changes OAuth providers\n- All data the victim creates after OAuth login is accessible to the attacker\n- The victim has no indication their account was pre-staged\n- Affects every OAuth provider configured in Authorizer (Google, GitHub, Facebook, Apple, LinkedIn, Twitter, Discord, Twitch, Roblox, Microsoft)\n\n## Suggested Fix\n\nBefore linking an OAuth identity to an existing account, verify that the existing account\u0027s email is already verified:\n\n    existingUser, err := h.StorageProvider.GetUserByEmail(ctx, refs.StringValue(user.Email))\n    if err == nil {\n        // Account exists. Only link if email is already verified.\n        if existingUser.EmailVerifiedAt == nil {\n            // Email not verified by original owner. Do NOT link.\n            // Either: reject the login, or create a new separate account,\n            // or delete the unverified account and create a fresh one for the OAuth user.\n        }\n    }\n\nAdditionally, when linking a new OAuth identity, invalidate any existing password on the account or require the user to re-authenticate via the original method.\n\n## Credit\nKoda Reef",
  "id": "GHSA-29rf-f4vv-pvq6",
  "modified": "2026-08-14T15:36:15Z",
  "published": "2026-08-14T15:36:15Z",
  "references": [
    {
      "type": "WEB",
      "url": "https://github.com/authorizerdev/authorizer/security/advisories/GHSA-29rf-f4vv-pvq6"
    },
    {
      "type": "WEB",
      "url": "https://github.com/authorizerdev/authorizer/commit/66fe488fd2a4e7acf1e517334344d5e8f3ddd296"
    },
    {
      "type": "PACKAGE",
      "url": "https://github.com/authorizerdev/authorizer"
    },
    {
      "type": "WEB",
      "url": "https://github.com/authorizerdev/authorizer/releases/tag/2.4.0-rc.16"
    }
  ],
  "schema_version": "1.4.0",
  "severity": [
    {
      "score": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:P/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N",
      "type": "CVSS_V4"
    }
  ],
  "summary": "Authorizer: Zero-click account takeover via OAuth identity linking to unverified email accounts"
}

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.

Detection rules are retrieved from Rulezet.

The MITRE ATT&CK techniques below are AI-generated suggestions, inferred from the description of the vulnerability by the CIRCL/vulnerability-attack-technique-classification-roberta-base model, served locally by ML-Gateway. They have not been verified by an analyst and are provided for guidance only.
The approach is described in our paper Mapping CVEs to MITRE ATT&CK Techniques: A Curated Gold-Set Classifier and the Limits of LLM-Assisted Label Expansion.