/

CVE-2026-88880

Severity:

HIGH RISK

(8.6)

Open Redirect (CWE-601) in Renovate GitLab Pagination

CVE-2026-88880

Awaiting Analysis

Vulnerability type

Category:

Open Redirect

CWE:

CWE-601

Impact:

Credential Exposure

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

Renovate versions before 44.11.3 do not sufficiently validate destinations provided in GitLab pagination Link headers. A compromised or malicious GitLab server may redirect credential-bearing requests to attacker-controlled infrastructure, potentially exposing authentication credentials.

Affected context

The issue occurs when Renovate communicates with a GitLab server and follows pagination links supplied through HTTP Link headers. Deployments may be at risk when the GitLab endpoint or an intermediary can return attacker-chosen pagination destinations and Renovate includes authentication material while following those links.

Exploit preconditions

  • Attack surface: HTTP communication between Renovate and a GitLab API - Required access: Control of, or the ability to influence responses from,

the GitLab server being queried

  • Trigger point: Processing and following a paginated response's Link header - Why exploitation is possible: An unvalidated destination can cause a

credential-bearing request to leave the intended GitLab trust boundary

Why this happens

The underlying bug is insufficient validation of redirect destinations supplied through pagination metadata. The client treats a server-provided Link target as trusted without adequately constraining its scheme, host, or relationship to the expected GitLab endpoint. This behavior maps to CWE-601, where an attacker-controlled redirect can move a request to an unintended destination and expose security-sensitive request data.

Potential impact

  • Authentication credentials may be disclosed to attacker-controlled infrastructure
  • Access tokens or other request headers could be captured
  • Unauthorized access to repositories or automation resources could result
  • A compromised GitLab service could facilitate broader credential-based attacks

General mitigation guidance

Mitigation generally focuses on treating pagination and redirect destinations as untrusted input. Secure-by-default client behavior can restrict followed links to approved schemes, hosts, and expected GitLab origins, while avoiding forwarding credentials across trust boundaries. Redirect handling can also minimize sensitive headers, use narrowly scoped credentials, and apply least privilege to tokens and service accounts. Network egress controls and explicit trust policies can provide additional containment.

How teams detect issues like this

Detection approaches can include static analysis and code review checks for HTTP clients that follow Link or redirect destinations without origin validation. Reviewers can look for unconditional forwarding of Authorization or cookie headers, unrestricted cross-host requests, and missing allowlists for API endpoints. Runtime monitoring can flag credential-bearing requests to unexpected domains, while CI policy enforcement can require redirect validation and sensitive-header stripping. Automated security review tools may help identify these patterns.

Sources

SourceLink
github.comgithub.com/renovatebot/renovate/security/advisories/GHSA-9hmg-9h89-jhmx
vulncheck.comvulncheck.com/advisories/renovate-before-44.11.3-credential-exfiltration-via-link-header

Frequently asked

questions

It can be exploitable through network communication when an attacker controls

or compromises the GitLab server that Renovate contacts and can influence its

pagination responses.

The attacker generally needs control of the GitLab endpoint, or another

position that allows modification of the relevant HTTP response. Direct

access to the Renovate host is not necessarily required.

Validate pagination destinations against an explicit set of trusted GitLab

origins and avoid forwarding credentials when a request crosses that trust

boundary.

Treat all redirect and Link header destinations as untrusted, enforce

origin-aware HTTP client policies, use narrowly scoped credentials, and

monitor egress for unexpected credential-bearing requests.

Find out what's already
exploitable in your codebase.