Tools AI Risk Radar ai-incident-0138
Incident record
MaxKB fixes prompt-driven command execution and sandbox escape
- Severity
- Critical
- Status
- Patched
- Type
- Prompt Injection
- Target
- MaxKB through 2.10.3-lts in the maintainer advisory
- Actor
- researcher
- CVE
- CVE-2026-77521
What happened
A tool-enabled MaxKB assistant could run shell commands after receiving malicious instructions through chat or retrieved content. The maintainer’s 2 September advisory scores the flaw 10.0 Critical under CVSS 3.1 and fixes it in 2.10.5-lts.
Lasso’s later investigation also tested 2.10.4-lts. On a direct installation, commands ran as the application user; its official-container demonstration escaped the application sandbox to container root. The researchers tested their own installations.
Sources
- Maintainer advisorygithub.com/1Panel-dev/MaxKB/security/advisories/GHSA-f36j-f3…
- Lasso researchwww.lasso.security/blog/cve-2026-77521-maxkbypass---from-pro…
One record from the AI Risk Radar, maintained by the Security Desk.Data: CSV · JSON ·RSS · CC BY 4.0 with attribution to YFarmX.Tracker updated · 13 October 2026