DRUPAL-CONTRIB-2026-268: Composer/drupal/ai vulnerability

Q: What conditions are required for exploitation?

An attacker must be able to supply untrusted sources to the AI module and cause an interaction that produces rendered HTML. The issue involves insufficient validation of links under those conditions.

Q: What is the potential impact?

Sensitive information may be exfiltrated to external sites through links that are not sufficiently validated.

Q: Which deployments are most exposed?

Deployments that allow untrusted sources to be used in AI-module interactions that render HTML are exposed to the described attack path. The provided information does not establish whether any particular default configuration permits this.

Q: What can be done while patching is unavailable?

Restrict or avoid use of untrusted sources in AI-module interactions that produce rendered HTML, since both are required for the described exploitation path.

Published Oct 28, 2026

Updated

This module enables you to use AI on the website. It also enables you to interact with AI providers.

The module does not sufficiently validate links under certain conditions, potentially allowing the exfiltration of sensitive information to external sites.

This vulnerability is mitigated by the fact that an attacker must use untrusted sources when interacting with the AI module in a way that produces rendered HTML.

Affected Software

2 affected componentsFixes available

composer/drupal/ai<1.4.11

1.4.11

composer/drupal/ai>=1.5.0<1.5.2

1.5.2

Remediation

Recommended actions to resolve this vulnerability, in priority order.

Upgrade
Upgrade composer/drupal/ai to a version that resolves this vulnerability.
Fixed in 1.4.11
Upgrade
Upgrade composer/drupal/ai to a version that resolves this vulnerability.
Fixed in 1.5.2

Event History

Oct 28, 2026

Advisory Published

via OSV·05:09 PM

Data Sourced

via OSV·05:09 PM

DescriptionAffected Software

Frequently Asked Questions

What conditions are required for exploitation?

An attacker must be able to supply untrusted sources to the AI module and cause an interaction that produces rendered HTML. The issue involves insufficient validation of links under those conditions.

What is the potential impact?

Sensitive information may be exfiltrated to external sites through links that are not sufficiently validated.

Which deployments are most exposed?

Deployments that allow untrusted sources to be used in AI-module interactions that render HTML are exposed to the described attack path. The provided information does not establish whether any particular default configuration permits this.

What can be done while patching is unavailable?

Restrict or avoid use of untrusted sources in AI-module interactions that produce rendered HTML, since both are required for the described exploitation path.

DRUPAL-CONTRIB-2026-268: Composer/drupal/ai vulnerability

Affected Software

Remediation

Event History

Frequently Asked Questions

What conditions are required for exploitation?

What is the potential impact?

Which deployments are most exposed?

What can be done while patching is unavailable?

Company

Resources

Contact

Frequently Asked Questions

What conditions are required for exploitation?

What is the potential impact?

Which deployments are most exposed?

What can be done while patching is unavailable?