Skip to main content

DeepTutor EUVDEUVD-2026-91125

| CVE-2026-51880 CRITICAL
Path Traversal (CWE-22)
2026-10-01 mitre GHSA-45p5-w2r8-mrfv
9.8
CVSS 3.1 · Vendor: mitre
View

Attack ChainAIDerived

Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain

  1. Access
    Reach live tutorbot WebSocket endpoint
  2. Delivery
    Send EditFileTool request with absolute/traversal path
  3. Exploit
    Write or edit file outside bot workspace
  4. Execution
    Overwrite sensitive config or code
  5. Impact
    Achieve code execution on host

Vulnerability Assessment

Weakness
CWE
OWASP Top 10:2025
A01:2025 - Broken Access Control
ASVS 5.0
V5 - File Handling Path canonicalization, traversal prevention
MITRE ATT&CK

2 techniques across 2 tactics: Discovery, Collection

The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up.

Exploitation Conditions

Exploitation requires that the DeepTutor 1.4.0 'live tutorbot' WebSocket interface is running and network-reachable by the attacker; the EditFileTool must be invocable through that tool layer (its default behavior in this version). Per CVSS PR:N/UI:N, no authentication or user interaction is needed on that interface, so any client able to open the WebSocket can drive the write. The primary LIMITING factor is deployment exposure: instances that do not run the live tutorbot server, or that restrict the WebSocket to localhost/trusted networks, are not remotely exploitable. No KEV-confirmed active exploitation, so risk is concentrated on internet-exposed instances.

Impact

Signals conflict sharply and must be weighed together. CVSS 3.1 rates this 9.8 Critical (AV:N/AC:L/PR:N/UI:N, all impacts High) and SSVC marks it Automatable=yes with Total technical impact, which argues for high priority. However, EPSS is only 0.14% (3rd percentile), indicating very low modeled probability of widespread exploitation, and the vulnerability is NOT in CISA KEV - so there is no evidence of active exploitation in the wild. A public exploit does appear to exist (SSVC Exploitation=poc, plus a referenced gist), so weaponization is trivial for anyone who targets an exposed instance. The dominant mitigating factor is reach: DeepTutor is a niche open-source AI tutoring project, not enterprise-ubiquitous software, so the realistic attack surface is limited to the small number of operators who actually run the live tutorbot WebSocket interface and expose it to untrusted networks. Net: a genuine high-severity flaw for anyone running the affected component, but the low EPSS and niche footprint make this targeted/opportunistic rather than a mass-exploitation emergency. Note also that NVD metadata is weak - CPE and EUVD list 'n/a' placeholders and the vendor (HKUDS) is unlabeled - so affected-version enumeration should be verified against the GitHub issue rather than NVD.

Exploit ScenarioDerived

An attacker reaches an exposed DeepTutor 1.4.0 live tutorbot WebSocket endpoint without authenticating and sends a crafted EditFileTool request specifying an absolute or traversal path outside the bot workspace. Because the tool layer honors the attacker-controlled path, the attacker writes or overwrites arbitrary files on the host - e.g., a startup script, SSH key, or application code - which can escalate to remote code execution. Publicly available exploit code exists (referenced gist), lowering the skill barrier.

Act Now Prioritize within 24-48h. Why: CRITICAL severity.
No patch listed yet 0.1% exploitation probability (EPSS)

Get an e-mail when this CVE is exploited, patched or re-scored. Free account.

Alert me

AnalysisAIv1 – Oct 05, 2026

Unauthenticated remote clients can escape the intended bot workspace and write or overwrite arbitrary absolute filesystem paths through DeepTutor 1.4.0's live 'tutorbot' WebSocket interface, because the EditFileTool.execute path-traversal flaw (CWE-22) fails to contain caller-supplied paths to the workspace root; with CVSS 3.1 AV:N/AC:L/PR:N/UI:N and total confidentiality/integrity/availability impact, this rates 9.8 and can corrupt configuration, application code, or authentication material. Publicly available exploit code exists (SSVC Exploitation: poc, Automatable: yes) and the issue is tracked upstream as HKUDS/DeepTutor issue 517 with a public proof-of-concept gist, but there is no KEV-confirmed active exploitation and EPSS is low at 0.14% (3rd percentile), so the practical risk is concentrated on instances whose tutorbot WebSocket is internet-reachable rather than restricted to localhost or a trusted network.

RemediationAI

No vendor-released patch has been identified at the time of analysis: no fixed version is specified in the advisory data, and the references point to upstream issue https://github.com/HKUDS/DeepTutor/issues/517 and a public PoC gist (https://gist.github.com/Ro1ME/b5ea6ac856a593e1d20042eb72c00fae) rather than to a tagged release, so treat any upstream commit as an unverified fix until the maintainer publishes a version. Until then, the highest-value control is exposure reduction: bind the live tutorbot WebSocket to 127.0.0.1 and reach it only through an authenticated reverse proxy or SSH tunnel, or block the port at the perimeter - the trade-off is that remote/legitimate tutorbot clients lose direct access and must be re-routed. Add an authentication and origin-check layer in front of the WebSocket and reject non-loopback peers, accepting the added operational complexity and latency. Run DeepTutor under a dedicated low-privilege account inside a container with the application directory and any secrets volume mounted read-only and with a writable scratch directory that is the only bind-mounted path, which converts an arbitrary-write into a contained write but can break features that legitimately edit project files. Apply filesystem hardening such as AppArmor/SELinux or systemd ProtectSystem=strict, ReadWritePaths limited to the workspace, and no write access to /etc, /root, ~/.ssh, or the Python site-packages directory, so that even a successful traversal cannot persist code or keys; verify no symbolic links from the workspace to sensitive locations exist. Finally, monitor DeepTutor release notes and the GitHub issue for an official patched version, and in the interim watch for anomalous file modifications outside the workspace by the service account as a detection signal.

Affected VersionsENISA

n/a n/a

Affects your stack

Find out whether the stacks and servers you run are on an affected version.

Checking this CVE against your own stacks and servers is part of the Basic plan and up.

Technical ContextAI · show details

DeepTutor is an agent-style tutoring framework in which an LLM tool layer exposes filesystem operations such as EditFileTool.execute to a conversational backend; the flaw is a classic CWE-22 improper limitation of a pathname to a restricted directory, i.e., the tool resolves attacker-influenced path strings without canonicalizing them and verifying that the result remains beneath the configured bot workspace, so absolute paths (and traversal sequences) are honored verbatim. The delivery channel is the live tutorbot WebSocket interface, which accepts tool-invocation messages and, per the assessed CVSS vector (PR:N, UI:N), requires neither authentication nor user interaction to drive the write - meaning any network client that can open the socket can reach the file-write primitive. The CPE string supplied is effectively unpopulated (n/a:n/a), and EUVD likewise lists no concrete version data, so product identification rests on the vendor repository HKUDS/DeepTutor and the issue reference; the SSVC decision point 'Technical Impact: total' reflects that an arbitrary-write primitive with root-adjacent or service-level file access can subvert the host rather than merely read data. Note that the 9.8 base score describes worst-case impact on a reachable interface, not the probability that any given deployment is exposed.

Know when Deeptutor gets its next vulnerability

  • Watch products and get an email when something new is published
  • Check this CVE against your own stack and servers
  • Exploit and PoC references, threat intelligence from 22 sources

No credit card · 7-day full trial

Severity by source

Vendor (mitre) PRIMARY 9.8 CRITICAL
AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
vuln.today AI 9.8 CRITICAL

Unauthenticated remote caller abuses the WebSocket EditFileTool to read/write/overwrite arbitrary host files, so AV:N/AC:L/PR:N/UI:N with High confidentiality, integrity, and availability via potential code overwrite.

3.1 AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H
4.0 AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N

Primary rating from Vendor (mitre).

CVSS VectorVendor: mitre

Attack Vector
Network
Attack Complexity
Low
Privileges Required
None
User Interaction
None
Scope
Unchanged
Confidentiality
High
Integrity
High
Availability
High

Lifecycle Timeline

3
CVSS changed
Oct 05, 2026 - 17:22 NVD
9.8 (CRITICAL)
CVE Published
Oct 01, 2026 - 00:00 cve.org
UNKNOWN (no severity yet)
CVE Published
Oct 01, 2026 - 00:00 cve.org
CRITICAL 9.8

DescriptionCVE.org

deeptutor 1.4.0 contains a path traversal issue in EditFileTool.execute. Through the live tutorbot WebSocket interface, a remote caller can induce the tool layer to write or edit absolute paths outside the intended bot workspace.

Executive ViewAI

Why It Matters

DeepTutor 1.4.0, an AI tutoring platform exposing a live 'tutorbot' WebSocket interface, lets any unauthenticated remote client write to or overwrite arbitrary files anywhere on the host filesystem, which exposes configuration data, application code, and authentication material to tampering or destruction. Any organization running an internet-reachable DeepTutor instance should treat this as a high-priority exposure, because a single unauthenticated request can undermine the integrity and availability of the platform and anything else on that host. Leadership should care less about the low exploit-prediction score and more about the fact that no vendor fix currently exists while a public proof-of-concept is already circulating.

Business Impact: High
Compensating Controls

No vendor-released patch identified at time of analysis. Interim mitigations: restrict the tutorbot WebSocket to localhost or a trusted network and block external reachability at the edge; enforce authenticated reverse-proxy access in front of the service; run DeepTutor under a dedicated low-privilege account inside a container or chroot with a read-only filesystem view outside the bot workspace; monitor file integrity on configuration, code, and authentication paths for unauthorized modifications.

Threat Intelligence Overview

EPSS Score
0.1%
Top 97% exploitable
Prevalence
Rare
obscure open-source research AI tutoring project, negligible production footprint
Data Quality
6 sources
Confidence: High

How this decision was reached

The engine reads each signal below and applies a fixed list of rules from top to bottom. The first rule that matches sets the recommended action – like a statement, every line is shown.

SignalValue
Severity CRITICAL
CISA KEV No
Public POC No
EPSS 0.1%
Patch available No
Active threat No
Deciding rule
CRITICAL severity
What every action level means
Emergency Immediate action required. Criteria: CRITICAL + in CISA KEV (actively exploited) without patch, OR CRITICAL + EPSS > 30% without patch, OR CRITICAL + active threat signals without patch.
Act Now Prioritize within 24-48h. Criteria: in KEV with patch available, OR CRITICAL + public POC, OR HIGH + EPSS > 10%, OR CRITICAL + active threat, OR any CRITICAL severity.
This Week Schedule remediation this week. Criteria: HIGH severity (CVSS 7.0-8.9). Patch within normal change window.
This Month Plan within 30 days. Criteria: MEDIUM severity (CVSS 4.0-6.9), lower exploitation probability. Include in regular patch cycle.
Monitor Track and reassess. Criteria: LOW severity (CVSS 0.1-3.9), minimal exploitation risk. No immediate action needed.

SSVC AssessmentENISA

SSVC Decision
Scheduled
Exploitation
Poc
System Exposure· derived
Open
Automatable
Yes
Human Impact
High
Technical Impact
Total

Priority Score

49
Low Medium High Critical
KEV: 0
EPSS: 0
CVSS: +49
PoC: 0
Weaponized: 0
Threat intel: 0
Gov advisory: 0

Attack Technique

Path traversal exploits occur when applications use user-controlled input to construct file system paths without proper validation.

This site uses cookies essential for authentication and security. No tracking or analytics cookies are used. Privacy Policy