Python
CVE-2026-33081
MEDIUM
Vulnerability Assessment
| Weakness |
CWE
OWASP Top 10:2025
A01:2025 - Broken Access Control
ASVS 5.0
V4 - API and Web Service URL allowlisting, SSRF prevention
MITRE ATT&CK
2 techniques across 2 tactics: Initial Access, Credential Access The technique list with ATT&CK links and the kill-chain view is part of the Basic plan and up. |
| Impact | The CVSS v3.1 score of 5.8 (Medium) reflects a network-accessible attack vector with low complexity and no privilege requirement, but indicates no confidentiality impact and only integrity impact via state changes on internal services. However, this score underestimates real-world risk in several ways: First, the proof-of-concept is fully functional and publicly available, demonstrating reliable exploitation. Second, the feature is opt-in (security.allowDownload must be explicitly enabled), limiting deployment scope but creating high risk for deployments that do enable it—such as document processing, screenshot, or archival services that intentionally enable download functionality. Third, while EPSS data is not provided in the intelligence, the blind SSRF nature means attackers cannot directly exfiltrate data but can perform reconnaissance, trigger administrative endpoints, access cloud metadata services (AWS IMDSv1 at 169.254.169.254), or interact with internal APIs. The attack complexity is very low because it requires only hosting a simple HTML page with a JavaScript redirect—no sophisticated payload engineering. The lack of authentication requirement (CWE-918 and CVSS vector confirm PR:N) and the fact that this bypasses the intended validateDownloadURL() security control elevate practical risk beyond the numeric score. Organizations running PinchTab with download enabled should prioritize remediation despite the medium CVSS rating. |
| Exploit ScenarioDerived | An attacker registers a domain or obtains access to a CDN capable of hosting arbitrary HTML content, then creates a simple HTML page containing JavaScript that waits 1.5 seconds before redirecting to http://127.0.0.1:1337/increment (or any internal service endpoint). The attacker then sends a single HTTP request to the target PinchTab instance with /download?url=https://attacker-controlled-domain.com, which passes initial validation because the attacker's domain is external and publicly routable. PinchTab launches a headless Chromium browser, loads the attacker's page, validation completes, and the browser then executes the JavaScript redirect to the internal service, triggering state changes such as incrementing counters, modifying configuration, or accessing cloud metadata endpoints—all originating from the PinchTab host's network context and thus bypassing firewall rules that would block direct external access. A proof-of-concept script demonstrating this exact attack flow is publicly available on the GitHub advisory, making reliable exploitation trivial. |
Get an e-mail when this CVE is exploited, patched or re-scored. Free account.
Alert meAnalysisAIv1 – Mar 18, 2026
PinchTab contains a Server-Side Request Forgery (SSRF) vulnerability in its /download endpoint that allows unauthenticated attackers to bypass URL validation and cause the embedded Chromium browser to make requests to internal network services. The vulnerability affects PinchTab versions 0.7.x and 0.8.x when the security.allowDownload setting is enabled (disabled by default), and exploits a validation gap where only the initial user-supplied URL is checked while subsequent browser-initiated requests (redirects, JavaScript navigations, resource fetches) bypass this protection entirely. Although the attacker cannot receive response bodies from internal services (blind SSRF), they can trigger state-changing endpoints on localhost or private network addresses reachable from the PinchTab host, with a proof-of-concept publicly available demonstrating counter increments on internal services.
RemediationAI
Upgrade PinchTab to a patched version released by the project maintainers (check https://github.com/pinchtab/pinchtab/releases for the latest security update). The primary fix involves applying network validation not only to the initial user-supplied URL but to all subsequent requests issued by the Chromium browser context, including redirects, navigations, and resource fetches. Until a patch is available, organizations should disable the security.allowDownload feature entirely (set to false) unless download functionality is absolutely required. If download functionality must remain enabled, implement network-level mitigations by restricting outbound traffic from the PinchTab process to external networks only, blocking all connections to loopback addresses (127.0.0.0/8) and private RFC 1918 ranges (10.0.0.0/8, 172.16.0.0/12, 192.168.0.0/16) at the firewall or via Linux network namespace policies (iptables, nftables). Additionally, run PinchTab in a restrictive container or virtual machine with explicit egress policies to prevent the browser from reaching internal services.
Affected Packages
Source: OSV (Open Source Vulnerabilities)
Affects your stack
Affects 1 package in go
Checking this CVE against your own stacks and servers is part of the Basic plan and up.
Technical ContextAI · show details
The vulnerability exists in the PinchTab download handler (pkg:go/github.com_pinchtab_pinchtab), which relies on a validateDownloadURL() function to filter requests to private and loopback addresses during the initial validation phase. However, this security function operates only once at request intake, before the headless Chromium browser instance is spawned and begins processing the attacker-controlled page. Modern web browsers are designed to execute JavaScript, follow redirects, load resources, and issue fetch requests—all asynchronous operations that occur after the initial page load. The root cause, classified under CWE-918 (Server-Side Request Forgery), stems from incomplete request validation: the application implements network boundary enforcement at the application layer for user input but fails to enforce the same policy at the browser context layer for all outbound requests. This creates a fundamental mismatch between the validation scope and the actual request surface, allowing attackers to use JavaScript or HTML meta-refresh redirects to pivot from an attacker-controlled external domain to internal addresses (127.0.0.1, private RFC 1918 ranges) after validation has already passed.
Know when Python 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
AV:N/AC:L/PR:N/UI:N/S:C/C:N/I:L/A:N
Primary rating from GitHub Advisory.
CVSS VectorGitHub Advisory
Lifecycle Timeline
2DescriptionGitHub Advisory
The /download endpoint validates only the initial URL provided by the user using validateDownloadURL() to prevent requests to internal or private network addresses.
Exploitation requires \security.allowDownload=true, which is disabled by default.
However, pages loaded by the embedded Chromium browser can trigger additional browser-side requests (for example, JavaScript redirects, navigations, or resource requests) after the initial validation step.
Because the validation is only applied to the initial URL and not to subsequent browser-issued request targets, an attacker-controlled page can cause the browser to issue requests to internal network services reachable from the PinchTab host.
This results in a blind Server-Side Request Forgery (SSRF) condition in which internal-only services may be accessed and state-changing endpoints may be triggered without returning the response body to the attacker.
Steps to Reproduce:
Environment Setup Target: PinchTab server (tested on v0.8.x, v0.7.x) Attacker-controlled server: Publicly accessible (e.g., via ngrok) attacker.py Internal service: Runs on the same host as PinchTab and is not externally accessible internal_service.py
1. Start a Local Internal Service (Victim Side)
Run a simple HTTP service bound to localhost: internal_service.py
python internal_service.py
#Example behavior of internal_service.py: #Listens on 127.0.0.1:1337 #Exposes endpoint /increment #Increments a counter and logs requests
#Expected output when accessed: #COUNTER INCREMENTED: 1 #COUNTER INCREMENTED: 2
2. Host an Attacker-Controlled Page (Attacker side)
Deploy a malicious HTML page that redirects to the internal service: attacker.py
<html> <body> <script> setTimeout(function(){ window.location = "http://127.0.0.1:1337/increment"; }, 1500); </script> </body> </html> Host this page on a publicly accessible server (e.g., using ngrok): https://fcb8-180-149-93-3.ngrok-free.app
3. Trigger the Vulnerable Endpoint (Attacker side)
Send a request to the PinchTab /download endpoint:
curl "http://[server-ip]:9867/download?url=https://fcb8-180-149-93-3.ngrok-free.app"
If a server token is configured, the request must include valid authentication.
4. Observe Server-Side Request to Localhost
When PinchTab processes the request:
- It launches a headless Chromium instance
- The browser loads the attacker-controlled page
- JavaScript executes within the browser
- The browser redirects to: http://127.0.0.1:1337/increment
5. Verify the Impact
Check the output of internal_service.py: <img width="718" height="156" alt="proof" src="https://github.com/user-attachments/assets/cf00e3e6-71c6-44ae-83b0-ed819f19ee9a" />
COUNTER INCREMENTED: 1 <img width="718" height="282" alt="proof_incremented" src="https://github.com/user-attachments/assets/98281b8e-221b-4e76-a10b-1b2335d08c61" />
This confirms that the request originated from the PinchTab host and that an attacker can successfully access localhost-only internal services via the browser, despite the initial URL validation.
Impact
This vulnerability allows an attacker to bypass the /download URL validation and cause the embedded Chromium browser to make requests to internal network services. By hosting a page that performs a redirect after the initial validation, an attacker can force the browser to access resources such as 127.0.0.1 or other private network addresses reachable from the PinchTab host.
Although the response is not returned to the attacker (blind SSRF), this behavior can still be used to interact with internal services and trigger state-changing endpoints. In environments where sensitive services or cloud metadata endpoints are accessible from the host, this could lead to more serious security impact.
Mitigation
Apply the same URL safety policy to every browser-issued request in the /download flow, not only the initial user-supplied URL, and block requests to loopback, private, link-local, and other non-public network ranges inside the Chromium browser context.
Executive ViewAI
CVE-2026-33081 presents moderate security risk, a server-side request forgery vulnerability rated CVSS 5.8. Users could be tricked into performing unintended actions. Organizations should apply vendor updates when available and monitor for exploitation.
Apply vendor patch. If patching is delayed: implement egress filtering, block metadata endpoints.
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
Server-Side Request Forgery exploits applications that fetch remote resources based on user-supplied URLs.
Affected Products (CPE)GitHub Advisory
Related Vulnerabilities in URL Validation Bypass in HTTP Clients
View all 2Wazuh SIEM platform versions 4.4.0 through 4.9.0 contain an unsafe deserialization vulnerability in the DistributedAPI t
BentoML version 1.4.2 and earlier contains an unauthenticated remote code execution vulnerability through insecure deser
pgAdmin 4 contains critical remote code execution vulnerabilities in the Query Tool download and Cloud Deployment endpoi
The renderLocalView function in render/views.py in graphite-web in Graphite 0.9.5 through 0.9.10 uses the pickle Python
BentoML is a Python library for building online serving systems optimized for AI apps and model inference. Rated critica
OpenSSL before 0.9.8za, 1.0.0 before 1.0.0m, and 1.0.1 before 1.0.1h does not properly restrict processing of ChangeCiph
pyLoad download manager version prior to 0.5.0b3.dev77 exposes the Flask SECRET_KEY through an unauthenticated endpoint.
Langflow (a visual LLM pipeline builder) contains a critical unauthenticated code execution vulnerability (CVE-2026-3301
In Mercurial before 4.1.3, "hg serve --stdio" allows remote authenticated users to launch the Python debugger, and conse
Unauthenticated remote code execution affects Kestra OSS (the open-source event-driven orchestration platform) prior to
Unauthenticated remote code execution in Marimo ≤0.20.4 allows attackers to execute arbitrary system commands via the `/
pyLoad is the free and open-source Download Manager written in pure Python. Rated medium severity (CVSS 5.3), this vulne
Same weakness CWE-918 – Server-Side Request Forgery (SSRF)
View allVendor StatusVendor
SUSE
Severity: Medium| Product | Status |
|---|---|
| openSUSE Leap 15.6 | Fixed |
| SUSE Linux Enterprise Module for Package Hub 15 SP5 | Fixed |
| SUSE Linux Enterprise Module for Package Hub 15 SP6 | Fixed |
| openSUSE Leap 15.5 | Fixed |
GHSA-qwxp-6qf9-wr4m
NVD
CVE.org
EPSS
VulDB
Exploit-DB