/

CVE-2026-92916

Severity:

HIGH RISK

(7.5)

Information Exposure (CWE-200) in Grav Clockwork Profiler

CVE-2026-92916

Deferred

Vulnerability type

Category:

Information Exposure

CWE:

CWE-200

Impact:

Sensitive Data Exposure and Potential Account Takeover

Severity:

HIGH RISK

(7.5)

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

Grav can expose its Clockwork profiler endpoint without authentication when the non-default debugger is enabled. An attacker who can reach the endpoint may retrieve sensitive request history, including session cookies, submitted credentials, configuration data, and stored secrets.

Affected context

The issue occurs in Grav deployments where the debugger is enabled and the Clockwork profiling endpoint is reachable through the web server, CDN, or another network path. Stored profiler records may contain request cookies, authorization headers, request bodies, system and plugin configuration, and operator-managed secrets, depending on application behavior and profiler settings.

Exploit preconditions

  • Attack surface: The web-accessible Clockwork profiler endpoint.
  • Required access: Network reachability to the endpoint; normal application

authentication is not necessarily required.

  • Trigger point: A request targeting a path containing /__clockwork/

while the debugger is enabled.

  • Why exploitation is possible: Request handling does not consistently

enforce authentication, local-origin restrictions, or an equivalent

profiler access control before returning stored history.

Why this happens

The underlying flaw is an access-control failure that exposes diagnostic data through an unauthenticated endpoint, mapping primarily to CWE-200. The bootstrap path routes matching profiler requests to the debugger without requiring a verified user, trusted client origin, or valid profiler token. Because profiler records retain sensitive request metadata and configuration values, insufficient endpoint authorization becomes a broad information disclosure issue. Inadequate redaction of nested form fields, cookies, and credential-related headers further increases the sensitivity of the exposed records.

Potential impact

  • An attacker may obtain active session identifiers and impersonate users.
  • An attacker could access credentials submitted through application forms.
  • An attacker may disclose SMTP credentials, API keys, license keys, and other secrets.
  • An attacker could use authorization or API-token headers to access connected services.
  • An attacker may gain operational intelligence about the site's system and plugins.
  • Exposure of administrator session data could lead to account takeover and content modification.

General mitigation guidance

Mitigation generally focuses on removing unauthenticated access to diagnostic

interfaces and minimizing the data retained by profiling systems.

Common defensive patterns include:

  • Disable debugging and profiling features in production environments unless they are explicitly required.
  • Restrict diagnostic endpoints to local or explicitly trusted requests.
  • Require a strong, separately managed secret or equivalent authentication before serving profiler data.
  • Block diagnostic paths at the web server, reverse proxy, or CDN when they are not needed.
  • Strip cookies, authorization headers, API tokens, passwords, and other secrets before records are stored.
  • Apply recursive redaction to nested request bodies rather than filtering only top-level fields.
  • Limit profiler retention, pagination scope, and access to stored diagnostic history.
  • Run the application with least privilege and rotate exposed credentials or session material when disclosure is suspected.

How teams detect issues like this

Organizations can identify this class of issue through:

  • Dynamic checks that request diagnostic paths without authentication and verify that no profiler data is returned.
  • Configuration reviews for enabled production debugging or profiling features.
  • Static analysis and code review searches for request handlers that dispatch diagnostic endpoints before authorization.
  • Tests confirming that cookies, passwords, authorization headers, API tokens, and secrets are recursively redacted.
  • Web-server and CDN policy checks ensuring diagnostic paths are denied or limited to trusted sources.
  • Monitoring for unusual requests to profiler endpoints and unexpected downloads of historical diagnostic records.
  • Automated security review tools that flag exposed debug routes and missing access-control checks.

Sources

SourceLink
github.comgithub.com/getgrav/grav/security/advisories/GHSA-q3ff-cj6v-rr5g
vulncheck.comvulncheck.com/advisories/grav-through-2.0.21-unauthenticated-information-disclosure-via-clockwork

Frequently asked

questions

Yes, when the debugger is enabled and the profiler endpoint is reachable over the

network. The attacker generally does not need to authenticate to the exposed

endpoint.

The key requirement is network reachability to the diagnostic endpoint. Access to

an ordinary Grav account may not be required because the affected request path

does not reliably perform user or token authorization.

Disable production debugging and profiling where they are unnecessary. If

diagnostics are required, protect the endpoint with strong access controls,

trusted-origin restrictions, and comprehensive redaction before data is stored

or returned.

Treat debugging endpoints as sensitive administrative interfaces, enforce

authorization before dispatch, test unauthenticated access in CI, minimize

diagnostic retention, and prevent secrets and session material from entering

profiler records.

Find out what's already
exploitable in your codebase.