/

CVE-2026-28213

Severity:

CRITICAL RISK

(9.8)

Sensitive Information Exposure in Authentication Flows

CVE-2026-28213

Analyzed

Vulnerability type

Category:

Sensitive Information Exposure in Authentication Flows

CWE:

CWE-200, CWE-640

Impact:

Account Takeover

Severity:

CRITICAL RISK

(9.8)

CVSS version:

3.1

Catch this CVE while code is being written

Detect vulnerable code directly in pull requests, so CVEs are fixed before merge.

Summary

EverShop's pre-2.1.1 "Forgot Password" API exposes the password reset token directly in its response when a target email address is submitted.

An attacker who knows or guesses a user's email can obtain a valid reset token and immediately take over that user's account without access to their inbox.

Affected context

This vulnerability affects EverShop deployments running versions prior to 2.1.1 that expose the public "Forgot Password" endpoint.

The issue appears when the application generates a password reset token and returns it in the API response instead of delivering it exclusively through a trusted out-of-band channel such as email.

Any environment where external users can invoke the password reset API and where email addresses can be guessed or enumerated is at risk.

Exploit preconditions

  • Attack surface: Public-facing "Forgot Password" API endpoint.
  • Required access: Ability to send HTTP requests to the password reset endpoint; no prior authentication is required.
  • Trigger point: Submitting a password reset request for a chosen email address via the vulnerable endpoint.
  • Why exploitation is possible: The server includes the password reset token in the API response payload, effectively disclosing a secret that should only be available through a separate, trusted channel. This bypasses the intended control that only the true account owner (with mailbox access) can use the token.

Why this happens

The core issue is a breakdown in the design of the password reset workflow, where a sensitive token is exposed directly to the requester instead of being kept confidential and delivered out-of-band.

This maps to CWE-200: Exposure of Sensitive Information to an Unauthorized Actor, because the reset token is a credential-like secret that grants control over the account.

It also aligns with CWE-640: Weak Password Recovery Mechanism for Forgotten Password, as the reset procedure fails to properly verify that the requester controls the associated email account and relies on a mechanism that leaks the recovery artifact to any caller.

Abstractly, the insecure pattern is: treating a recovery token as a regular API field rather than as a secret that must only be shared through a secure delivery channel.

Potential impact

  • An attacker can obtain valid password reset tokens for arbitrary accounts by supplying their email addresses.
  • Full account takeover for affected users, including access to personal data and order history.
  • Potential unauthorized purchases, refunds, or changes to shipping and billing details.
  • Reputational damage and loss of customer trust due to compromised accounts.
  • Possible regulatory or contractual issues if personal data is exposed through compromised accounts.

General mitigation guidance

Effective mitigation focuses on redesigning the password reset flow so that reset tokens are treated as secrets and never returned via normal API responses.

Recommended defensive patterns include:

  • Ensuring password reset tokens are only delivered through a trusted out-of-band channel (e.g., emailed as a link) and never included in the response body or logs.
  • Treating reset tokens as time-limited, single-use secrets stored securely server-side and validated only during the reset confirmation step.
  • Returning generic, non-enumerable responses from password reset initiation (e.g., "If this email exists, we sent instructions") to avoid confirming account existence.
  • Applying rate limiting and abuse detection to password reset endpoints to reduce automated attacks.
  • Designing authentication and recovery mechanisms with clear trust boundaries and minimizing the exposure of any secret required to authenticate or reset credentials.

How teams detect issues like this

Organizations can identify this class of vulnerability using a combination of review and testing approaches:

  • Code review: Examine password reset and account recovery handlers for any logic that returns reset tokens, verification codes, or similar secrets in HTTP responses or logs.
  • Static analysis: Configure rules to flag code paths that serialize sensitive fields (such as token, resetToken, verificationCode) into API responses or log entries.
  • Dynamic testing: Exercise the "Forgot Password" flow with a controlled test account and inspect all responses to confirm that no reset tokens or recovery codes are exposed.
  • Security testing in CI/CD: Include automated checks for common authentication and recovery mistakes, and require manual review of changes touching auth, session, or password reset code paths.

Sources

SourceLink
github.comgithub.com/evershopcommerce/evershop/releases/tag/v2.1.1
github.comgithub.com/evershopcommerce/evershop/security/advisories/GHSA-cg73-g723-39jw

Frequently asked

questions

Yes. Anyone with network access to the public EverShop "Forgot Password" endpoint can exploit the issue without needing an existing account or prior authentication.

The attacker only needs to be able to send HTTP requests to the password reset API and know or guess the victim's email address. No access to the victim's mailbox is required, because the token is disclosed directly in the API response.

The most robust fix is to ensure that password reset tokens are never returned to clients via normal API responses, and are instead sent out-of-band (such as via email) while being stored securely server-side as time-limited, single-use secrets.

Treat all authentication and recovery artifacts (tokens, codes, links) as secrets, apply secure design reviews for authentication flows, and enforce coding guidelines and automated checks that prohibit exposing such secrets in responses, logs, or client-visible fields.

Find out what's already
exploitable in your codebase.