CVE-2026-105783: Joplin: Web Clipper CORS flaw leaks permanent API token
HIGHPoC AVAILABLEJoplin Desktop before 3.7.13 has a flaw in its opt-in Web Clipper server. It answers cross-origin requests from any website, so a malicious page can start the pairing flow and, if the user clicks Approve on a generic prompt that names no origin, receive a permanent API token. That token gives lasting read and write access to notes, folders, tags, resources and master keys, which matters because notes often hold credentials, prompts, API keys and research. Joplin is not an AI/ML framework, so this is only tangentially AI-related: exposure is limited to teams that keep AI project notes or secrets in it. Exploitation needs user interaction and has high complexity. NVD lists no public exploit, no Nuclei template and no KEV listing, and the EPSS score is low at 0.00145, so this is not an emergency. Upgrade Joplin Desktop to 3.7.13 or later, and disable the Web Clipper service where it isn't needed.
What is the risk?
Moderate-to-high severity (CVSS 8.0) with low practical exposure. Exploitation requires the Web Clipper server to be enabled (opt-in), a victim visiting an attacker-controlled page, and the victim approving an ambiguous dialog (UI:R, AC:H). EPSS is 0.00145, there is no known public exploit and no KEV entry. The scope change and high confidentiality and integrity impact come from the persistent token, which survives the browsing session. The AI/ML link is indirect: it applies only where Joplin stores AI-related secrets or documentation.
How does the attack unfold?
What systems are affected?
| Package | Ecosystem | Vulnerable Range | Patched |
|---|---|---|---|
| joplin | — | — | No patch |
Do you use joplin? You're affected.
How severe is it?
What is the attack surface?
What should I do?
1 step1) Upgrade Joplin Desktop to 3.7.13 or later (fix in PR #16276, commit 27ae5c0). 2) Disable the Web Clipper service in Joplin settings if you don't use it. 3) Don't approve Web Clipper pairing prompts you didn't start yourself. 4) If you suspect compromise, disable and re-enable the clipper to rotate the token, re-key E2EE if master keys may have been exposed, and rotate any secrets stored in notes. 5) Detection: look for unexpected local connections to the clipper port (default 41184) from browser processes, unexpected pairing dialogs, and unfamiliar note edits. 6) Keep secrets in a dedicated secrets manager, not in note-taking apps.
What does CISA's SSVC say?
Source: CISA Vulnrichment (SSVC v2.0). Decision based on the CISA Coordinator decision tree.
How is it classified?
Which compliance frameworks are affected?
This CVE is relevant to:
Frequently Asked Questions
What is CVE-2026-105783?
Joplin Desktop before 3.7.13 has a flaw in its opt-in Web Clipper server. It answers cross-origin requests from any website, so a malicious page can start the pairing flow and, if the user clicks Approve on a generic prompt that names no origin, receive a permanent API token. That token gives lasting read and write access to notes, folders, tags, resources and master keys, which matters because notes often hold credentials, prompts, API keys and research. Joplin is not an AI/ML framework, so this is only tangentially AI-related: exposure is limited to teams that keep AI project notes or secrets in it. Exploitation needs user interaction and has high complexity. NVD lists no public exploit, no Nuclei template and no KEV listing, and the EPSS score is low at 0.00145, so this is not an emergency. Upgrade Joplin Desktop to 3.7.13 or later, and disable the Web Clipper service where it isn't needed.
Is CVE-2026-105783 actively exploited?
Proof-of-concept exploit code is publicly available for CVE-2026-105783, increasing the risk of exploitation.
How to fix CVE-2026-105783?
1) Upgrade Joplin Desktop to 3.7.13 or later (fix in PR #16276, commit 27ae5c0). 2) Disable the Web Clipper service in Joplin settings if you don't use it. 3) Don't approve Web Clipper pairing prompts you didn't start yourself. 4) If you suspect compromise, disable and re-enable the clipper to rotate the token, re-key E2EE if master keys may have been exposed, and rotate any secrets stored in notes. 5) Detection: look for unexpected local connections to the clipper port (default 41184) from browser processes, unexpected pairing dialogs, and unfamiliar note edits. 6) Keep secrets in a dedicated secrets manager, not in note-taking apps.
What systems are affected by CVE-2026-105783?
This vulnerability affects the following AI/ML architecture patterns: Developer workstations, Knowledge management and note-taking tools, Local desktop applications with REST APIs.
What is the CVSS score for CVE-2026-105783?
CVE-2026-105783 has a CVSS v3.1 base score of 8.0 (HIGH). The EPSS exploitation probability is 0.14%.
What is the AI security impact?
Affected AI Architectures
MITRE ATLAS Techniques
AML.T0036Data from Information RepositoriesAML.T0078Drive-by CompromiseAML.T0091.000Application Access TokenCompliance Controls Affected
What are the technical details?
Original Advisory
Joplin is an open source note-taking and to-do application that organises notes and lists into notebooks. Prior to 3.7.13, when Joplin Desktop is running with the opt-in Web Clipper server enabled, the server in packages/lib/ClipperServer.ts sends Access-Control-Allow-Origin: * and allows an arbitrary website to call POST /auth and GET /auth/check because the pairing endpoints do not reject HTTP or HTTPS origins. The desktop confirmation dialog does not identify the requesting origin, so a victim who approves the generic prompt authorizes the attacking page, which then receives the permanent API token. The token provides ongoing read and write access to notes, folders, tags, resources, and master keys. This issue is fixed in version 3.7.13.
Exploitation Scenario
An attacker lures a researcher or engineer to a web page, for example a fake AI paper or tool landing page. Script on the page makes cross-origin calls to POST /auth and GET /auth/check on the local Web Clipper server. The victim sees a generic Joplin authorization prompt that doesn't show the requesting origin and approves it, assuming it is a legitimate clipping request. The page receives the permanent API token and exfiltrates the notebooks, including stored LLM API keys and internal AI project notes. It can then keep reading the notes or modify them for as long as the token remains valid.
Weaknesses (CWE)
CWE-346Origin Validation ErrorPrimaryCWE-352Cross-Site Request Forgery (CSRF)PrimaryCWE-346Origin Validation ErrorCWE-352Cross-Site Request Forgery (CSRF)CWE-346 — Origin Validation Error: The product does not properly verify that the source of data or communication is valid.
Source: MITRE CWE corpus.
CVSS Vector
CVSS:3.1/AV:N/AC:H/PR:N/UI:R/S:C/C:H/I:H/A:NReferences
- github.com/laurent22/joplin/commit/27ae5c0d404ee47ad24862cce917dda6eb2930b7x_refsource_MISC
- github.com/laurent22/joplin/pull/16276x_refsource_MISC
- github.com/laurent22/joplin/releases/tag/v3.7.13x_refsource_MISC
- github.com/laurent22/joplin/security/advisories/GHSA-9728-v7ww-mxjvx_refsource_CONFIRM
Timeline
Related Vulnerabilities
CVE-2023-376510.0MLflow: path traversal allows arbitrary file read
Same attack type: Data LeakageCVE-2025-512010.0smolagents: sandbox escape enables unauthenticated RCE
Same attack type: Data LeakageCVE-2025-282810.0LangChain RequestsToolkit: SSRF exposes cloud metadata
Same attack type: Auth BypassCVE-2025-5376710.0Azure OpenAI: SSRF EoP, no auth required (CVSS 10)
Same attack type: Auth BypassCVE-2026-2185810.0n8n: Input Validation flaw enables exploitation
Same attack type: Auth Bypass