Skip to main content

JoySafeter EUVDEUVD-2026-104683

| CVE-2026-109443 HIGH
Improper Isolation or Compartmentalization (CWE-653)
2026-10-20 VulnCheck GHSA-qmmc-575c-v2h8
8.7
CVSS 4.0 · Vendor: VulnCheck
View

Attack ChainAIDerived

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

  1. Access
    Self-register low-privilege account
  2. Delivery
    Open WebSocket chat.start session
  3. Exploit
    Select Copilot standard mode
  4. Execution
    Invoke unsandboxed execute tool
  5. Persist
    subprocess.run shell=True on host
  6. Impact
    Read cross-tenant data and secrets

Vulnerability Assessment

Weakness
CWE
OWASP Top 10:2025
A06:2025 - Insecure Design
ASVS 5.0
V15 - Secure Coding and Architecture Component isolation
Exploitation Conditions

Exploitation requires an authenticated low-privilege account (PR:L), but the barrier is minimal because the application permits self-registration; the attacker must then use the WebSocket chat.start extension, select Copilot 'standard mode', and trigger the execute tool, which runs subprocess.run with shell=True without sandboxing. Concrete prerequisites: open/self-service registration enabled, standard mode available to the account, and network reachability to the WebSocket endpoint. Limiting factors are the need for a valid (self-created) session and a reachable instance; no user interaction from a victim and no high privileges are required, so default internet-facing deployments with open registration are directly exposed.

Impact

This is a credible operational priority, not a paper tiger. The CVSS 4.0 base vector (AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:H, score 8.7 High) and the description agree: a network-reachable, low-complexity attack requiring only low privileges and no user interaction yields full confidentiality, integrity, and availability impact via host operating-system command execution. The low-privilege requirement (PR:L) is effectively trivial because the description states attackers can self-register, so the authentication barrier offers little real defense where open registration is enabled. Publicly available exploit code exists (per VulnCheck/Lasso references), which raises near-term risk, but this is NOT the same as active exploitation: there is no CISA KEV listing and no EPSS score was provided, so exploitation probability cannot be quantified from the input. CWE-653 (improper isolation/compartmentalization) is consistent with the reported cross-tenant data and server-secret exposure. The one tempering factor is deployment footprint: this is an obscure JD open-source project (jd-opensource/joysafeter) at version 1.0.0, so despite the high severity and available POC, real-world blast radius is limited by low prevalence. Signals are internally consistent - no contradictions between vector and description - the main missing data are EPSS and any vendor-confirmed patched version.

Exploit ScenarioDerived

An attacker self-registers a standard user account on an internet-facing JoySafeter instance, opens a WebSocket chat.start session, and selects Copilot 'standard mode'. They then prompt the agent to use its unsandboxed execute tool, causing the backend to call subprocess.run with shell=True and run arbitrary OS commands on the host to read other tenants' data and server secrets. A public exploit/POC is referenced, lowering the skill needed; the low attack complexity (AC:L) and no required user interaction make this straightforward once an account exists.

This Week Schedule remediation this week. Why: HIGH severity (CVSS 7.0-8.9).
No patch listed yet 0.7% exploitation probability (EPSS)

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

Alert me

AnalysisAIv1 – Oct 20, 2026

Authenticated remote code execution in JoySafeter through 1.0.0 allows any self-registered low-privilege user to run host operating-system commands by selecting Copilot standard mode and invoking its unsandboxed execute tool over the WebSocket chat.start extension. Successful exploitation can expose other tenants' data and server secrets because the agent calls subprocess.run with shell=True outside a sandbox, and publicly available exploit code exists. The assessed CVSS 3.1 vector requires only PR:L with no user interaction, so internet-facing deployments that allow open registration are directly exposed and warrant operational priority.

RemediationAI

No vendor-released patch or fixed version is identified in the provided data, so treat JoySafeter through 1.0.0 as unpatched at time of analysis and apply a vendor fix as soon as one is published while tracking the VulnCheck advisory at https://www.vulncheck.com/advisories/joysafeter-through-1.0.0-authenticated-rce-via-copilot-standard-mode-shell. Immediate compensating controls are to disable self-service registration or require administrative approval, which removes the attacker's easiest path but affects onboarding; disable Copilot standard mode or the execute tool for low-privilege accounts, which breaks agent command-execution features; run agent and backend workers in a locked-down container or VM with no host secrets, a read-only filesystem, dropped Linux capabilities, seccomp or AppArmor, and a dedicated low-privilege service account, which adds operational overhead and may break legitimate tool use; restrict the WebSocket chat.start endpoint to trusted networks or authenticated SSO groups, which reduces exposure but limits remote users; and monitor for unexpected subprocess or shell activity from the JoySafeter service account. Do not rely on authentication alone because self-registration makes the PR:L prerequisite trivial to obtain.

Affected VersionsENISA

JoySafeter 0 ≤1.0.0

Affected Packages

euvd joysafeter no fix

Source: OSV (Open Source Vulnerabilities)

Affects your stack

Affects 1 package in euvd

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

Technical ContextAI · show details

JoySafeter is a multi-tenant AI copilot and agent platform whose backend command executor at backend/app/core/agent/backends/utils/command_executor.py line 110 uses subprocess.run with shell=True, while backend/app/services/copilot_service.py line 106 exposes the execute tool to Copilot standard mode. The WebSocket chat.start extension lets an authenticated user drive the agent, and standard mode lacks the isolation that should separate tool execution from the host. CWE-653, Improper Isolation or Compartmentalization, is the root cause: the agent runtime, tool dispatcher, and operating-system shell share the same trust boundary, so a low-privilege user's request reaches the host shell with the service account's privileges. The CPE cpe:2.3:a:jd-opensource:joysafeter:*:*:*:*:*:*:*:* confirms the affected product family, and the resulting host access creates cross-tenant data and secret disclosure risk.

Know when Joysafeter 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 (VulnCheck) PRIMARY 8.7 HIGH
CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N
vuln.today AI 8.8 HIGH

Network-reachable (AV:N), low complexity (AC:L), self-registered low-privilege account (PR:L), no interaction (UI:N); unsandboxed host command execution yields full C/I/A.

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

Primary rating from Vendor (VulnCheck).

CVSS VectorVendor: VulnCheck

Attack Vector
Network
Attack Complexity
Low
Attack Requirements
None
Privileges Required
Low
User Interaction
None
Vulnerable system: Confidentiality
High
Vulnerable system: Integrity
High
Vulnerable system: Availability
High
Subsequent system: Confidentiality
None
Subsequent system: Integrity
None
Subsequent system: Availability
None

DescriptionCVE.org

JoySafeter through 1.0.0 contains a remote code execution vulnerability that allows authenticated users to run operating system commands via an unsandboxed execute tool in Copilot standard mode. Attackers can self-register, select standard mode in the WebSocket chat.start extension, and have the agent invoke subprocess.run with shell=True on the host to access other tenants' data and server secrets.

Technical POC Analysis

Source of evidence: Lasso Security vulnerability database entry (no vendor advisory, no patch diff, no CVE record text supplied).

Full technical POC analysis - root cause, step-by-step exploitation and patch breakdown - is part of the Pro and Business plans.

Executive ViewAI

Why It Matters

JoySafeter through 1.0.0 carries a high-severity authenticated code-execution flaw that can let a low-privilege self-registered user compromise the underlying host and potentially expose other tenants' data and server secrets. Because publicly available exploit code exists and internet-facing deployments with open registration are directly exposed, leadership should treat this as a priority containment and tenant-trust issue.

Business Impact: High
Compensating Controls

Restrict or disable self-registration, disable Copilot standard mode and the unsandboxed execute tool, tightly control WebSocket chat.start access, place JoySafeter behind VPN or allowlists, run it with least privilege in a segmented container or host, restrict outbound access, separate tenant data and secrets, rotate exposed secrets, and monitor registration, WebSocket, and host command logs.

Threat Intelligence Overview

EPSS Score
0.7%
Top 47% exploitable
Prevalence
Low
niche JD open-source project, early 1.0.0 release, limited production adoption
Exploit
POC
Public exploit code exists
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 HIGH
CISA KEV No
Public POC Yes
EPSS 0.7%
Patch available No
Active threat No
Deciding rule
HIGH severity (CVSS 7.0-8.9)
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
No
Human Impact
High
Technical Impact
Total

Priority Score

65
Low Medium High Critical
KEV: 0
EPSS: +1
CVSS: +44
PoC: +20
Weaponized: 0
Threat intel: 0
Gov advisory: 0

Attack Technique

Remote Code Execution represents the critical moment when an attacker successfully runs arbitrary code on a target system without physical access.

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