CVE-2026-79575: yfexam-exam v2.0 Auth Bypass Vulnerability

CVE-2026-79575 Overview

CVE-2026-79575 is a cryptographic weakness in yfexam-exam v2.0, an online examination platform. The application derives its JSON Web Token (JWT) signing secret from a combination of the username and the current month rather than a random server-side key. This predictable derivation allows attackers to brute-force the signing secret and forge valid authentication tokens. The weakness maps to [CWE-330] (Use of Insufficiently Random Values) and is exploitable over the network without authentication or user interaction.

Critical Impact

An unauthenticated remote attacker can recover the JWT signing secret and forge tokens to impersonate any user, resulting in full confidentiality compromise of session-protected resources.

Affected Products

• yfexam-exam v2.0

Discovery Timeline

• 2026-09-08 - CVE-2026-79575 published to NVD
• 2026-09-09 - Last updated in NVD database

Technical Details for CVE-2026-79575

Vulnerability Analysis

yfexam-exam v2.0 issues JWTs to authenticated users to maintain session state. Instead of signing these tokens with a cryptographically random secret held server-side, the application constructs the signing key deterministically from the username and the current calendar month. Because both inputs are low-entropy and partially known to an attacker, the secret space collapses to a search domain small enough for brute force on commodity hardware.

Once the secret is recovered, an attacker can mint arbitrary tokens with any user identity or claim set. The server accepts these forged tokens as legitimate, granting access to protected endpoints and data belonging to the impersonated account.

Root Cause

The root cause is deterministic key derivation from predictable inputs. Secure JWT implementations use a high-entropy, secret key generated once at server initialization and stored outside the application logic. Deriving the key from a username exposes it to any attacker who knows or can guess the username, and rotating monthly does not restore entropy because the current month is public information.

Attack Vector

The attack proceeds over the network against the authentication endpoint or any endpoint that validates JWTs. An attacker enumerates or guesses a valid username, computes candidate secrets by combining the username with the current month value, and validates each candidate against a captured token signature. After recovering the secret, the attacker forges a new JWT claiming the target identity and submits it in the Authorization header. Refer to the GitHub proof-of-concept for exploitation details.

No verified exploit code is reproduced here; consult the linked proof-of-concept for the concrete derivation logic.

Detection Methods for CVE-2026-79575

Indicators of Compromise

• JWTs presented to the application with valid signatures but originating from IP addresses or user-agents inconsistent with the claimed user's normal activity.
• Authentication events for accounts that show no corresponding login flow in application logs.
• Sudden access to administrative or high-privilege endpoints from previously inactive accounts.

Detection Strategies

• Inspect JWT iat and exp claims for tokens whose issuance times do not correlate with recorded login events on the server.
• Correlate token usage against the source IP history of each user account to surface impersonation.
• Enable verbose logging on the JWT validation layer to capture the full token header, payload, and requesting client metadata for forensic review.

Monitoring Recommendations

• Alert on repeated failed signature validations, which may indicate an attacker probing candidate secrets against a captured token.
• Monitor for high-volume authenticated requests without a preceding successful login event in the same session window.
• Track access patterns to sensitive exam-management endpoints and flag deviations from historical baselines.

How to Mitigate CVE-2026-79575

Immediate Actions Required

• Rotate the JWT signing configuration to a cryptographically random secret of at least 256 bits, generated once and stored in a secrets manager or environment variable.
• Invalidate all currently issued JWTs by changing the signing key, forcing re-authentication for every user.
• Restrict network access to the yfexam-exam application until a fixed configuration is deployed.

Patch Information

No vendor patch is referenced in the CVE record at the time of publication. Consult the GitHub proof-of-concept repository and the yfexam-exam project maintainers for remediation guidance.

Workarounds

• Replace the deterministic key derivation function with a call to a secure random generator and persist the resulting secret outside the codebase.
• Add short token lifetimes and server-side session revocation to limit the value of any forged token.
• Deploy a reverse proxy or web application firewall rule that requires additional client-side attestation on JWT-protected endpoints until the underlying code is fixed.

# Configuration example: generate a strong JWT secret and export it for the application
openssl rand -base64 64 > /etc/yfexam/jwt.secret
chmod 600 /etc/yfexam/jwt.secret
export JWT_SECRET="$(cat /etc/yfexam/jwt.secret)"