Live Exposure Defense: From CVE to Confirmed Exposure in 12 Hours – See more

New CVE Detected

CVE-2026-72713 – Arbitrary File Read via Path Traversal – XAgent (OpenBMB) before commit 26f2b6e

Be the first to know when new zero-days emerge:

Summary

CVE-2026-72713 is a path traversal vulnerability in OpenBMB’s XAgent, an open-source autonomous AI agent framework, affecting the /workspace/file endpoint of its XAgentServer component. An authenticated attacker can supply directory-traversal sequences in the file_name parameter to read arbitrary files on the host, including application secrets, database credentials, and system files outside the intended Docker sandbox. The vulnerability carries a high-severity rating (base score 8.7) and is remotely exploitable over the network with low attack complexity.

Technical details

  • Root cause: The POST /workspace/file endpoint (implemented in XAgentServer/application/routers/workspace.py) concatenates the user-supplied file_name form field to a base directory using os.path.join() without validating or sanitizing the input for ../ sequences or absolute paths.
  • Trigger conditions: An attacker submits a request to /workspace/file with file_name set to a traversal payload (e.g., ../../../../../../etc/passwd), along with a valid user_id, token, and interaction_id.
  • Attack vector: Network-based; authentication is required, but valid credentials are trivially obtainable — XAgent’s registration endpoint (POST /user/register) auto-approves new accounts without email verification when send_email=False (the default configuration), and the application also ships with a documented default credential pair (admin/xagent-admin).
  • Impact: Successful exploitation allows arbitrary file read on the underlying host, breaking the Docker sandbox isolation the tool relies on. Exposed data can include .env files with database credentials, JWT and SMTP secrets, SSH keys, and other system files. A related issue in the same report — missing user-scope filtering in get_interaction() — also permits cross-user access to interaction data (IDOR), compounding exposure risk.

Affected software

  • Vendor/Product: OpenBMB XAgent
  • Affected versions: All versions/commits prior to the fix commit 26f2b6edc75127af524f027c022b382967178e3a

Severity

  • CVSS Score: 8.7 (High)
  • CVSS 4.0 Vector: CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N
  • Attack Vector: Network

Mitigation and recommended actions

  • Immediate: Upgrade XAgent to a version that includes commit 26f2b6edc75127af524f027c022b382967178e3a (merged via pull request #432), which introduces path_security.py validation utilities (safe_child_path(), validate_plain_filename(), safe_upload_path(), safe_workspace_path()) to reject absolute paths, null bytes, backslashes, and .. sequences across file upload/download and workspace endpoints. The patch also removes the separate /register_new_tool endpoint, which previously allowed arbitrary code registration.
  • If immediate patching is not possible:
    • Restrict network exposure of the XAgentServer (default listens on 0.0.0.0:8090) to trusted networks only; do not expose it directly to the internet.
    • Disable self-service registration or enforce email verification (send_email=True) to prevent unauthenticated actors from easily obtaining accounts.
    • Change or remove the default admin/xagent-admin credentials immediately.
    • Run the service under a least-privilege account and container isolation that does not expose sensitive files (.env, credential stores, SSH keys) to the workspace filesystem path.
    • Monitor logs for requests to /workspace/file containing ../ sequences or unexpected absolute paths.

References

Are you exposed?

Get a free report of your organization’s exposure to this CVE and threat

How IONIX’s External Exposure Management Platform Detects and Validates
Zero-Days to Shrink MTTR

1

Map your entire attack surface (continously)

IONIX uses multi-factor discovery methods, including DNS analysis, certificate mapping, metadata inspection, and more, to automatically map every internet-facing asset across your environment. This includes cloud instances, third-party platforms, shadow IT, and even forgotten infrastructure that traditional tools miss.

2

Monitor for new CVEs

Dozens of threat intel feeds using agentic technology are continuously analyzed to detect the appearance of proof-of-concept code, exploit kits, and indicators of active targeting. IONIX goes further by applying AI to proactively evaluate whether emerging vulnerabilities are likely to be exploited, even before PoCs go public.

3

Identify Potential External Exposures

Not all CVEs matter. IONIX filters vulnerabilities by asking attacker-centric questions: Can it be reached from the internet? Does it require authentication? Is it being exploited in the wild? This dramatically reduces noise and focuses teams on threats that can actually be weaponized.

4

Create Safe, Scalable Exploit Validations

IONIX transforms real-world PoCs into safe, non-intrusive test payloads that can be run in production environments without disruption. These simulations are precisely targeted to the systems that are vulnerable, ensuring rapid validation without unnecessary load.

5

Execute Exploit Validations

By combining context about software stack, versioning, exposure status, and reachability, IONIX ensures that only the right payloads are executed against the right assets, maximizing efficiency and minimizing risk.

6

Drive Fast and Actionable Remediation

Results are routed through integrations with ticketing, SOAR, and SIEM tools. Issues are written in plain language, bundled into remediation clusters, and prioritized based on asset criticality, exploitability, and blast radius. This shortens mean time to remediation (MTTR) and empowers teams to act with confidence.

Are you exposed?

Get a free report of your organization’s exposure to this CVE and threat

Subscribe to Threat Center RSS

Copy/paste the link below into your preferred RSS reader or follow these instructions to subscribe to Slack alerts.

Get Real-Time CVE Alerts to Your Email

Be the first to know when new zero-days emerge