/

CVE-2026-82641

Severity:

HIGH RISK

(8.6)

Missing Authentication (CWE-306) in Keploy Agent API

CVE-2026-82641

Deferred

Vulnerability type

Category:

Missing Authentication

CWE:

CWE-306

Impact:

TLS Traffic Disclosure and Recording Manipulation

Severity:

HIGH RISK

(8.6)

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

Keploy versions 3.1.0 through 3.6.25 expose the agent control-plane HTTP server on all network interfaces without authentication. An attacker who can reach this service may retrieve TLS session key material and traffic data, potentially decrypting recorded communications, or manipulate recording sessions through control endpoints.

Affected context

The vulnerability occurs in Keploy agent deployments where the control-plane HTTP server is reachable beyond a trusted local boundary and does not require authentication. Risk is higher when network controls allow untrusted clients to access agent endpoints that expose TLS keylog data or control recording and mock-storage operations.

Exploit preconditions

  • Attack surface: The Keploy agent control-plane HTTP server and its HTTP endpoints
  • Required access: Network reachability to the server; no valid credentials are required
  • Trigger point: Requests to endpoints that expose keylog data or control recording sessions
  • Why exploitation is possible: The service binds to all interfaces while omitting an authentication check, allowing reachable clients to invoke sensitive operations

Why this happens

The underlying flaw is missing authentication on a network-facing control-plane service. Sensitive administrative endpoints are exposed without verifying the identity or authorization of the caller, which maps to CWE-306: Missing Authentication for Critical Function. Binding the service to all interfaces expands the reachable attack surface, while the absence of access controls allows unauthenticated requests to obtain TLS session key material or alter recording state.

Potential impact

  • Attackers may obtain NSS keylog lines associated with recorded TLS sessions
  • Recorded TLS traffic may be decrypted when corresponding traffic data is available
  • Sensitive application data carried in captured sessions could be disclosed
  • Recording sessions may be stopped or altered without authorization
  • Stored mocks or test artifacts could be manipulated
  • The exposed service could support further compromise of the hosting environment

General mitigation guidance

Mitigation generally focuses on reducing service exposure and enforcing authentication and authorization for every control-plane operation.

Common defensive patterns include:

  • Binding administrative services to a narrowly scoped interface rather than all interfaces by default
  • Requiring strong authentication and endpoint-level authorization for sensitive operations
  • Separating traffic-capture controls from ordinary application network paths
  • Restricting access with network policies, firewalls, service meshes, or equivalent boundary controls
  • Avoiding exposure of TLS key material through unauthenticated or broadly reachable interfaces
  • Running the agent with the least privileges and limiting access to captured traffic and stored artifacts
  • Applying secure-by-default configuration templates and treating control-plane endpoints as administrative functions

How teams detect issues like this

Detection can combine configuration review, application testing, and network monitoring.

  • Static analysis: Identify HTTP servers that bind administrative interfaces to wildcard addresses without authentication middleware.
  • Code review signals: Look for control-plane routes that return key material or mutate recording state without identity and authorization checks.
  • Configuration review: Check whether agent listeners are exposed beyond trusted local or management networks.
  • Dynamic testing: Verify that unauthenticated requests to sensitive endpoints are rejected and that authorization is enforced consistently.
  • Policy enforcement: Use deployment policies to prohibit publicly reachable administrative listeners and require network segmentation for control-plane services.
  • Monitoring: Alert on unexpected access to keylog, recording-control, and mock-storage endpoints.

Sources

SourceLink
github.comgithub.com/keploy/keploy/commit/a6257d2b3184b85eb30edad345464aa292297b83
github.comgithub.com/keploy/keploy/pull/4451
github.comgithub.com/keploy/keploy/releases/tag/v3.6.26
github.comgithub.com/keploy/keploy/security/advisories/GHSA-p79c-x224-cv8h
vulncheck.comvulncheck.com/advisories/keploy-3.1.0-through-3.6.25-unauthenticated-tls-key-exposure
github.comgithub.com/keploy/keploy/issues/4394

Frequently asked

questions

Yes, when the Keploy agent control-plane HTTP server is reachable over the network. Exploitation does not require a local shell or a valid account if the exposed endpoints accept unauthenticated requests.

The attacker needs network connectivity to the agent's control-plane listener. The required reachability may result from broad interface binding, permissive firewall rules, exposed container ports, or insufficient network segmentation.

Restrict the listener to a trusted interface or network, require authentication and authorization for every control-plane endpoint, and prevent unauthenticated access to TLS keylog and captured traffic data.

Treat administrative APIs as privileged services, apply secure-by-default network exposure settings, test authorization in automated security checks, and monitor for unexpected access to sensitive control-plane routes.

Find out what's already
exploitable in your codebase.