Organizations Face Supply Chain Risks as Docker Patches Critical Token Reassignment Flaw CVE-2026-105445

Docker has resolved a critical vulnerability in Docker Hub that permitted authenticated attackers to impersonate other users through personal access token reassignment. The flaw, tracked as CVE-2026-105445, allowed unauthorized access to container repositories, potentially compromising software supply chains by enabling image tampering under legitimate names.

What happened

The vulnerability existed in Docker Hub’s account-to-organization conversion feature, allowing attackers to transfer control of personal access tokens without rotating secrets. An attacker could create an account, generate a token, convert it to an organization, and designate a victim as owner. This process reassigned the token to the target user without their consent or knowledge of the transfer.

Consequently, the attacker retained the token secret while the registry linked it to the victim’s identity. This allowed the attacker to authenticate as the target user and perform operations within the token’s scope. Affected actions included pulling, pushing, or deleting images in personal repositories and accessible organizations, according to Docker’s security bulletin.

The flaw was introduced on 26 August 2024, when the conversion process began reassigning tokens without revoking secrets. It remained active until 30 September 2026, when Docker received a report and deployed a fix to production within four hours. The exposure window lasted approximately two years, affecting any token transferred during this period.

Why it matters

Docker rated the issue critical because it posed a software supply chain risk, enabling attackers to publish malicious images under trusted names. Although the flaw did not grant web access or allow changes to passwords or two-factor settings, it permitted full repository manipulation within token scopes. This could disrupt builds and expose downstream users to compromised container images.

Docker investigated 304,000 pushes made with transferred tokens and analyzed high-risk public images but found no evidence of exploitation. Affected users and organization owners were notified on 2 October, and transferred tokens were deactivated on 8 October, except those explicitly retained by owners. The incident highlights risks in credential-ownership handling during account transitions.

What happens next

Organisations should verify whether any personal access tokens were transferred during account conversions between August 2024 and September 2026. Teams must review repository access logs for unauthorised pull, push, or delete operations during this period. Docker recommends rotating all personal access tokens to ensure no compromised secrets remain active in production environments.

Patch priority

Teams usually prioritise updating systems with direct internet exposure to reduce immediate attack surfaces. Check whether any internal registries mirror Docker Hub images to assess potential contamination risks. Back up current configurations before applying updates to ensure recovery options remain available. Roll out fixes in stages, testing critical workloads first to confirm stability. Monitor authentication logs for unusual token usage patterns that might indicate ongoing exploitation attempts.

What to do and how to stay safe: Docker

• Audit all personal access tokens for transfers occurring between August 2024 and September 2026.
• Rotate any tokens that were reassigned during account-to-organization conversions to invalidate compromised secrets.
• Review repository access logs for unauthorised pull, push, or delete operations during the exposure window.
• Verify that no transferred tokens remain active unless explicitly retained by the original owner.

Step-by-step guide: Vulnerability Prioritization: Stop Fixing Everything and Start Fixing What Matters

General security guidance from the AppSec Scoop newsroom. It is not confirmed advice from the organisations named in this story.

Frequently asked questions

Did Docker find evidence that attackers exploited this vulnerability?

No, Docker found no evidence of exploitation after examining 304,000 pushes and analyzing high-risk public images.

Could attackers change victim passwords or two-factor settings using this flaw?

No, the vulnerability did not provide access to Docker Hub web settings, preventing changes to passwords or authentication methods.

When did the vulnerability exist and when was it fixed?

The flaw existed from 26 August 2024 until Docker deployed a fix on 30 September 2026, shortly after receiving a report.