CVE-2026-54520 AI Agent Automation: Workflow file step path traversal allows read and write outside the expected directory
Authenticated users can exploit path traversal in workflow steps to read or write files outside the intended directory, potentially compromising sensitive data or application integrity.
- Exploited in the wild
- Not known
- Login needed
- Any signed-in user
- Attackable
- Over the internet
- Victim action needed
- Not needed
- Fix
- None listed
What is CVE-2026-54520?
CVE-2026-54520 is a high-severity path traversal vulnerability in vmDeshpande ai-agent-automation, published on 17 Sep 2026. Any signed-in user can exploit it over the network, with no action from the victim.
Official description
AI Agent Automation is a modular AI agent workflow automation platform with schedulers, tools, and observability. Prior to 0.9.1, the executeStep file-step implementation in backend/src/agents/executor.js passes the user-controlled step.path value through path.resolve with process.cwd() and then uses the resulting path for read or write operations without checking that it remains in an approved workflow directory. An authenticated user who can create or modify workflow file steps can supply traversal segments to escape the intended workspace and read sensitive files or write and overwrite files accessible to the backend process, including application-adjacent files when process permissions allow. This issue is fixed in version 0.9.1.
Weakness: CWE-22
Which versions are affected by CVE-2026-54520?
How do I fix CVE-2026-54520?
No fix has been published for CVE-2026-54520 yet. Limit exposure of the affected software until one is available.
Upgrade, and resolve every user-supplied path against an allowed base folder before opening it. More about Path traversal.
Is CVE-2026-54520 being exploited?
There is no known exploitation of CVE-2026-54520 in the wild and no public exploit code.
Public exploit code: 1 proofs of concept on GitHub.
Sources
Also known as GHSA-cm8g-8jfq-887p.