Safeguard
Vulnerability Analysis•

CVE-2026-34568: Stored XSS in Blog Posts in CI4MS

CVE-2026-34568 is a critical stored XSS in CI4MS blog post content shown in several views. Upgrade ci4ms to 0.31.0.0 or later and review stored posts.

Safeguard Security Team
3 min read

CVE-2026-34568 is a critical stored cross-site scripting flaw in CI4MS, a CodeIgniter 4 based CMS. NVD scores it 9.1. Blog post content is stored without sanitizing and rendered unsafely in several views. Upgrade to ci4ms 0.31.0.0 or later. Review stored posts after patching.

What's affected

CI4MS (ci4-cms-erp ci4ms) versions before 0.31.0.0. The fixed release is 0.31.0.0.

Any site running an older build with the affected feature reachable to a signed in user is in scope.

How to check if you're exposed

Find the ci4ms version your site runs. If you deployed from a git checkout, git describe --tags in the project folder shows the nearest release tag. Otherwise look at the release or commit you installed from.

Anything older than 0.31.0.0 is in scope. The record names no workaround, so the version number is the test. The advisory linked under Sources lists the exact affected range.

What to do now

  1. Upgrade ci4ms to 0.31.0.0 or later and redeploy.
  2. Until you can, limit who holds accounts that can create or edit blog posts, and review those accounts.
  3. Look through existing blog posts for entries that contain script tags or event handler attributes, and clean them.
  4. Sign out active sessions for administrators after the upgrade.

Why it matters

The weakness class is CWE-79, cross-site scripting. Blog post content is stored server side and then rendered in multiple application views without output encoding.

Stored XSS is worse than the reflected kind because the script waits in the database and runs for every person who later opens the page. The CVSS vector has Scope Changed, which reflects that the script runs in the browser of someone else, often an administrator. The usual fix is output encoding on every place the stored value is rendered, plus input validation.

The record does not flag this CVE as listed in CISA's Known Exploited Vulnerabilities catalog, and the advisory I read does not say it is exploited.

Treat existing post content as untrusted until you have looked at it. Patching stops new bad input from rendering, but anything already stored in the database still needs a check or a clean up. Export the post bodies and search them for script tags and inline handlers.

Quick check

What should I check in existing content?

Search your stored posts for script tags and inline event handler attributes. Rich text editors can hide these, so compare what the editor shows with the stored HTML.

Sources

Last reviewed: 19 October 2026.

Never miss an update

Weekly insights on software supply chain security, delivered to your inbox.

Self-healing security runs on Safeguard.

Your first fix PR is minutes away.

No sales call required, even your agent can complete the purchase over MCP.