Server-Side Request Forgery (SSRF)
In a Server-Side Request Forgery (SSRF) attack, the attacker abuses functionality on the server to read or update internal resources. By supplying or modifying a URL that the server-side code processes, the attacker can reach internal services, read cloud metadata, and access systems that are not exposed to the internet.
SSRF has become one of the highest-impact vulnerability classes in modern cloud environments. Accessing the cloud metadata endpoint (http://169.254.169.254) from an SSRF can yield temporary IAM credentials, leading to full cloud account compromise. Even in non-cloud environments, SSRF enables port scanning of internal networks, accessing internal admin panels, and reaching databases or caches that trust requests from localhost.
Attackers use a wide range of bypass techniques to defeat SSRF filters: DNS rebinding, IP address encoding variations (decimal, octal, hex), URL parser inconsistencies, IPv6 representations, and redirect chains. Partial SSRF — where the attacker controls only part of the URL — can still be exploitable through open redirects or URL fragment abuse.
SSRF commonly appears in webhook integrations, URL preview features, PDF generators, image processors, and any functionality that fetches remote resources on behalf of the user.
This page collects SSRF research, bypass techniques, cloud exploitation methods, and real-world writeups from bug bounty programs and security research.
From OWASP
SSRF is a position, not a payload
The useful way to think about server-side request forgery is that you are not exploiting a parser — you are borrowing the server's network position. Everything the application server can reach and you cannot is now in scope: internal APIs with no authentication because they were "not exposed", admin interfaces bound to localhost, container orchestration endpoints, databases with permissive network ACLs, and the cloud metadata service that will hand out credentials to anything that asks from the right place.
That last one is why SSRF severity climbed so sharply in cloud environments. On IMDSv1 a single unauthenticated GET returns role credentials, which converts a URL-fetch feature into access to whatever that role can do. IMDSv2's requirement for a PUT with a token header broke the simplest version of that chain, and a substantial amount of the research here concerns whether a given SSRF primitive is expressive enough to satisfy it.
Finding SSRF means finding places the server fetches on your behalf, and they are rarely labeled as such: webhook configuration, URL preview and unfurling, PDF and screenshot generation, image proxies and resizers, file import by URL, XML parsers resolving external entities, and integrations that accept a callback address. Anything that renders HTML server-side is worth particular attention, because the fetch may happen several layers below the input you control.
Defenses fail in predictable ways, and the bypass literature is mostly a catalogue of the gap between how a blocklist parses a URL and how the HTTP client does. DNS rebinding defeats validate-then-fetch. Redirects move the final destination past the check. Alternate IP encodings, IPv6 mappings and unusual authority syntax confuse naive parsers. The controls that hold up are architectural rather than syntactic: resolve and validate the address immediately before connecting, refuse to follow redirects across trust boundaries, egress-filter the service, and require authentication on internal services instead of relying on their being unreachable.
| Date Added | Link | Excerpt |
|---|---|---|
| 2026-11-19 NEW 2026 | SSRF Filters Bypass Using Redirects intermediate Bug Bounty | This content likely details how to bypass Server-Side Request Forgery (SSRF) filters by leveraging redirects. Attackers might exploit vulnerabilities in how applications handle redirected requests to make the server fetch resources from unintended locations, potentially leading to data exposure or unauthorized access. The core idea revolves around crafting specific redirect chains or exploiting misconfigurations in the SSRF protection mechanisms to circumvent intended security controls. → osintteam.blog |
| 2026-11-19 NEW 2026 | Exploiting XXE to Perform SSRF Attacks: A Step-by-Step Walkthrough intermediate XXE | This article details how to exploit XML External Entity (XXE) vulnerabilities to perform Server-Side Request Forgery (SSRF) attacks. The walkthrough demonstrates how an "overly trusting" XML parser can be manipulated. The primary goal is to go beyond simply reading local files and instead trick the parser into leaking sensitive cloud credentials. The technique involves understanding how XXE can be leveraged to make the server issue requests to internal or external resources on behalf of the attacker. → infosecwriteups.com |
| 2026-11-19 NEW 2026 | Atomic Server unauthenticated full-response SSRF intermediate | Atomic Server has a vulnerability classified as an unauthenticated full-response Server-Side Request Forgery (SSRF). This flaw resides within the `/bookmark` endpoint of the Atomic-Server application. The SSRF vulnerability allows attackers to make requests from the server to arbitrary URLs, and the full response from these requests is returned to the attacker. |
| 2026-11-19 NEW 2026 | Outstatic GitHub proxy sends OAuth token to non-GitHub origins intermediate Secrets | Outstatic's GitHub REST proxy has a vulnerability where it may send an OAuth token to origins other than GitHub. This could expose sensitive user credentials to unintended recipients, potentially leading to unauthorized access or data breaches. The issue lies in how the proxy handles requests and token forwarding, allowing for the transmission of the token to non-GitHub domains. |
| 2026-11-19 NEW 2026 | Obot SSRF intermediate | Obot from 0.8.0 before 0.27.0 contains a server-side... |
| 2026-11-19 NEW 2026 | temporal-proxy gRPC listener exposes static credentials intermediate Secrets | A vulnerability has been identified in the temporal-proxy gRPC listener that exposes static credentials. Attackers who can send requests to this listener could potentially gain unauthorized access. The issue stems from the static nature of the credentials, making them predictable and exploitable. |
| 2026-11-18 NEW 2026 | Authorize.net CIM Pro for Magento unauthenticated SSRF news | Writeup detailing an unauthenticated Server-Side Request Forgery (SSRF) in the Rootways Authorize.net CIM Pro extension for Magento 2, specifically before version 4.1.5. This vulnerability, tracked as CVE-2026-130583, is present in the Apple Pay merchant validation controller, allowing attackers to leverage `file://`, `gopher://`, or `dict://` URLs to exfiltrate application credentials and encryption keys from `app/etc/env.php`, or to access internal services and cloud metadata endpoints. |
| 2026-11-18 NEW 2026 | N-able N-central SSRF news | N-able N-central versions prior to 2026.4.1 HF3 contain a Server-Side Request Forgery (SSRF) vulnerability. This flaw arises from insufficient validation within the software. The SSRF vulnerability allows for unauthorized access and potential manipulation of internal resources by an attacker. Users are advised to update to a patched version to mitigate this security risk. |
| 2026-11-18 NEW 2026 | GitHub Enterprise Server SSRF news | Advisory for GHSA-m75h-w4cr-gwx4 detailing a Server-Side Request Forgery in GitHub Enterprise Server. An authenticated user with repository write access could exploit this by sending requests to internal services via Slack webhook validity checks. This vulnerability, impacting versions prior to 3.23, was fixed in 3.22.5, 3.21.10, and 3.20.12. Exploitation necessitated Secret Protection with secret scanning and validity checks for partner patterns to be enabled. |
| 2026-11-18 NEW 2026 | Pwnedlabs: SSRF to Pwned intermediate | This Pwnedlabs challenge involved exploiting a Server-Side Request Forgery (SSRF) vulnerability on the `app.huge-logistics.com` corporate website. The lab provided only the URL, requiring manual exploration without initial credentials or hints. The objective was to find and leverage the SSRF vulnerability to gain unauthorized access or control. |
| 2026-11-18 NEW 2026 | Typecho SSRF news | Library for detecting and mitigating a Typecho SSRF vulnerability before 1.3.1-rc. This vulnerability, found in the XML-RPC pingback.ping method, leverages a Time-of-Check to Time-of-Use (TOCTOU) window by exploiting DNS rebinding. An unauthenticated attacker can exploit this by configuring a domain to present a safe IP during validation and then an internal IP during the outbound request, enabling blind SSRF to arbitrary internal hosts. |
| 2026-11-17 NEW 2026 | HackTheBox - Nimbus intermediate Python Talks | This content details a penetration test on a system called Nimbus. The process begins with an Nmap scan, followed by an examination of the Nimbus job scheduler. A healthcheck within the scheduler reveals sensitive information, specifically "aws.nimbus.htb," which results in an "InvalidClientToken" error. The focus is on identifying vulnerabilities through reconnaissance and initial exploitation attempts. |
| 2026-11-17 NEW 2026 | SeaweedFS SSRF in VolumeServer news | Writeup of SeaweedFS SSRF in VolumeServer, affecting versions prior to 4.42. This vulnerability in the `VolumeServer.FetchAndWriteNeedle` gRPC method allows unauthenticated attackers to bypass the fix for CVE-2026-73080. Exploitation involves crafting `replicas[].url` entries or an `azure_endpoint` to force the volume server to send requests with controlled tokens and bodies to internal services and metadata addresses, referencing GHSA-939f-xh28-2q48 and CVE-2026-129851. |
| 2026-11-17 NEW 2026 | DeepAudit SSRF news AI | Writeup on CVE-2026-129880 in DeepAudit, a server-side request forgery vulnerability affecting versions 2.0.0-beta.5 through 3.8.0. Registered users can exploit this by creating a malicious gitlab or gitea project pointing to a controlled host. Requesting branches of this project allows attackers to capture operator tokens and access internal hosts. |
| 2026-11-17 NEW 2026 | Frappe Framework SSRF news Python | Writeup of CVE-2026-130069, a server-side request forgery vulnerability in Frappe Framework before versions 15.123.0 and 16.50.0. Exploitable by authenticated users, including portal Website Users, through unescaped email templates in the `frappe.email.email_body.get_email_html` method. This allows injection of stylesheet link tags, causing Premailer to fetch internal or external hosts and expose their content. |
| 2026-11-17 NEW 2026 | crw-server SSRF news | Advisory for crw-server before 0.45.1 detailing a server-side request forgery (SSRF) vulnerability. Attackers can exploit this by manipulating `proxy` and `proxy_list` fields in `/v1/scrape` to reach internal hosts, including loopback and private network addresses, bypassing URL checks. References include GHSA-x79h-wjj6-9xcp and CVE-2026-130077. |
| 2026-11-16 NEW 2026 | Rustrak SSRF intermediate | Vulnerability writeup detailing Rustrak's server-side request forgery (SSRF) in versions before 0.26.0. This flaw, identified as CVE-2026-127093, allows authenticated users to target arbitrary loopback, private-network, or cloud metadata addresses via the `routing_override.url` parameter in POST requests to `/api/integrations/{id}/test`. The vulnerability enables attackers to leverage custom headers and redirects, and extract response bodies from non-2xx status codes. |
| 2026-11-16 NEW 2026 | OpenEMR SSRF intermediate | OpenEMR versions prior to 8.5.0 are vulnerable to a server-side request forgery (SSRF) flaw. This vulnerability allows an attacker to trick the server into making unintended requests to internal or external resources. Exploiting this could lead to unauthorized access, data breaches, or further compromise of the system. Users are strongly advised to update to version 8.5.0 or later to patch this security issue. |
| 2026-11-16 NEW 2026 | Post to Google My Business blind SSRF intermediate | Library for detecting blind SSRF in the Post to Google My Business WordPress plugin. Versions prior to 3.5.7 are vulnerable, allowing authenticated users with Contributor privileges to exploit an unvalidated `mbp_post_attachment` URL. This vulnerability, reachable via the preview REST route through a raw `getimagesize()` call, enables attackers to perform internal port scanning, probe internal services and cloud metadata endpoints, and trigger GET-based side effects on internal systems by bypassing network perimeter controls. |
| 2026-11-16 NEW 2026 | Kroki SSRF in BlockDiag family intermediate | Advisory for CVE-2026-129485 detailing a server-side request forgery (SSRF) vulnerability in Kroki versions prior to 0.36.0. This vulnerability, present in the BlockDiag family of engines, allows unauthenticated attackers to craft diagrams and submit them to specific endpoints, enabling them to query internal services, access cloud metadata, or read local image files from the server via node icon and background attributes. |
| 2026-11-16 NEW 2026 | OWASP #7: Server-Side Request Forgery (SSRF) — Trick Your Server beginner API Sec | OWASP Top 7 identifies Server-Side Request Forgery (SSRF) as a critical vulnerability. SSRF occurs when an attacker can coerce a server-side application into making an HTTP request to an arbitrary domain of the attacker's choosing. This allows attackers to interact with internal services, scan the internal network, or access sensitive data that the server might have access to but would otherwise be protected from external access. Developers should implement strict input validation and network access controls to mitigate SSRF risks. |
| 2026-11-15 NEW 2026 | HTB Sau: Write-up intermediate | This write-up details the process of compromising the HTB Sau machine. The initial steps involved enumeration using Nmap to identify open ports, followed by service version detection on those ports (22, 80, 8338, 55555). This systematic approach is the foundation for uncovering vulnerabilities and progressing towards full machine access. |
| 2026-11-15 NEW 2026 | Huginn SSRF news | Advisory for CVE-2026-127083 in Huginn, a server-side request forgery vulnerability, arises when OutboundProxy.configure! incorrectly applies egress proxying only to Faraday and Typhoeus. This flaw allows authenticated users to leverage JiraAgent or WeiboPublishAgent to direct HTTParty, open-uri, or Net::HTTP requests to internal services, with JiraAgent capable of returning response data as events. |
| 2026-11-15 NEW 2026 | IBM ContextForge MCP Gateway SSRF news AI | IBM ContextForge MCP Gateway versions 1.0.0 through 1.0.11 are vulnerable to a Server-Side Request Forgery (SSRF) flaw. This vulnerability allows an attacker to manipulate the gateway into making requests to arbitrary internal or external resources. Exploiting this flaw could lead to unauthorized access to sensitive information or compromise of the system. Users are advised to update to a patched version to mitigate this security risk. |
| 2026-11-15 NEW 2026 | SiYuan 3.8.10 SSRF in skill-install news | SiYuan versions prior to 3.8.10 have a vulnerability in the skill-install feature that allows for server-side request forgery (SSRF). This means an attacker could potentially trick the SiYuan server into making requests to internal or external resources it shouldn't access. This could lead to information disclosure or further compromise of the system. Users are advised to update to SiYuan 3.8.10 or later to patch this security flaw. |
| 2026-11-15 NEW 2026 | @utcp/http 1.1.14 SSRF news AI | Advisory for @utcp/http version before 1.1.14 detailing a Server-Side Request Forgery (SSRF) vulnerability. This issue enables remote manipulation of loopback services by exploiting a {host} placeholder to bypass security checks. Attackers can target internal services on agent machines by supplying host arguments like 127.0.0.1 or leveraging Streamable HTTP and SSE discovery. |
| 2026-11-14 NEW 2026 | From $1,000 to $10,000: How I Turned a "Limited" XXE Into a Critical Redis Session Store Exposure intermediate Bug Bounty XXE | This blog post details how a researcher exploited a seemingly limited XXE vulnerability to achieve a critical exposure of Redis session data. The author describes the process of escalating the initial XXE to gain access to sensitive session information, highlighting the steps taken to transition from a minor finding to a high-impact vulnerability. The title indicates a significant progression in the value of the findings, suggesting a substantial increase in the reported bug bounty payout. |
| 2026-11-14 NEW 2026 | Exploiting XXE to Perform SSRF Attacks: A Step-by-Step Walkthrough intermediate Burp XXE | This walkthrough details how to exploit XML External Entity (XXE) vulnerabilities to execute Server-Side Request Forgery (SSRF) attacks. The process involves crafting malicious XML payloads that leverage XXE to trick the application into making requests to internal or external resources on the attacker's behalf. By understanding how XXE can be abused, developers can better defend against these types of vulnerabilities. The guide provides practical steps for identifying and demonstrating XXE-based SSRF. |
| 2026-11-14 NEW 2026 | HTB Cohort: SSRF Filter Bypass to Pre-Auth RCE via a Marimo Notebook advanced RCE | Hack The Box Cohort showcased a complex attack chain leading to pre-authentication Remote Code Execution (RCE). The exploit began with an SSRF filter bypass utilizing IP encoding. This allowed access to an internal status leak, which in turn revealed a critical notebook RCE vulnerability. Finally, a PackageKit race condition was leveraged to achieve full system compromise. |
| 2026-11-14 NEW 2026 | The Bouncer Who Waved Everyone In When the Scanner Broke: A $3,500 GitLab SSRF intermediate Bug Bounty | A $3,500 GitLab SSRF vulnerability was discovered due to a flawed error handling mechanism. The scanner, designed to prevent DNS rebinding attacks, had a catch block that incorrectly handled errors, essentially allowing all traffic through when the scanner broke. This "fail-open" scenario created a security gap, as a correct DNS-rebinding defense was undone by this single oversight. The vulnerability highlights the importance of robust error handling in security implementations. |
| 2026-11-14 NEW 2026 | Vvveb incomplete SSRF fix allows authenticated SSRF news | Vvveb version 1.0.8.6 has a vulnerability where an incomplete Server-Side Request Forgery (SSRF) fix allows authenticated SSRF attacks. This means a logged-in user could exploit the flaw to make the server send requests to arbitrary locations, potentially leading to unauthorized access or further compromise of internal systems. The incomplete patch failed to fully address the underlying SSRF issue. |
| 2026-11-14 NEW 2026 | jDownloads 4.1 SSRF via external download URL field news | The jDownloads Joomla extension, in version 4.1, has a critical Server-Side Request Forgery (SSRF) vulnerability. This flaw arises from how the extension handles external download URLs. An attacker can exploit this by inputting a malicious URL into the external download URL field. This allows them to force the server to make requests to arbitrary internal or external resources, potentially leading to information disclosure, unauthorized access, or further compromise of the server. Users are advised to update their jDownloads installation to a patched version to mitigate this security risk. |
| 2026-11-14 NEW 2026 | Flowise = 3.1.4 server-side request forgery news AI | Advisory for GHSA-vqvj-w792-h4m6 detailing a server-side request forgery vulnerability in Flowise versions prior to 3.1.4. Authenticated attackers can exploit this by creating or updating chatflows/agentflows, bypassing Custom Function sandbox protections. This bypass is achieved by requiring a module subpath within an allowlisted package, such as `axios/dist/node/axios.cjs` or `node-fetch/lib/index.js`. The exploit allows attackers to issue arbitrary GET and POST requests via `/api/v1/node-custom-function` to internal network addresses or cloud metadata endpoints, potentially leading to network mapping, credential disclosure, and internal service state changes. |
| 2026-11-13 2026 | Synology DiskStation Credential API SSRF news | A severe server-side request forgery (SSRF) vulnerability has been discovered in Synology DiskStation's credential API. This flaw allows attackers to trick the DiskStation into making requests to arbitrary internal or external resources on their behalf. This could lead to unauthorized access to sensitive information or further compromise of internal networks. Exploiting this vulnerability does not require any authentication. |
| 2026-11-13 2026 | HashiCorp Vault integrated storage SSRF news Secrets | HashiCorp Vault and Vault Enterprise have a critical Server-Side Request Forgery (SSRF) vulnerability affecting integrated storage. This flaw allows an attacker with read access to sensitive information to potentially exploit the vulnerability. The issue arises when Vault processes specific requests with manipulated inputs, enabling the attacker to make unauthorized requests to arbitrary hosts. This could lead to the exposure of sensitive data or further compromise of the system. |
| 2026-11-13 2026 | Grafana Prometheus data source SSRF news | Grafana's Prometheus data source has a Server-Side Request Forgery (SSRF) vulnerability. The data source, when configured to use an HTTP proxy, incorrectly handles resource paths. This allows an attacker to manipulate the path to send arbitrary HTTP requests to internal or external servers, potentially exposing sensitive information or allowing for further exploitation of the Grafana server or other network resources. The vulnerability is present in specific versions of Grafana. |
| 2026-11-13 2026 | SOGo before 5.12.13 SSRF news | Advisory for SOGo before 5.12.13 detailing a server-side request forgery (SSRF) vulnerability. Authenticated users can exploit the Calendar/addWebCalendar endpoint by providing a crafted URL to force the server to connect to arbitrary internal host:port destinations. Attackers can then use the reload endpoint and analyze response variations to probe internal network details from the server's perspective. |
| 2026-11-13 2026 | code16/sharp SSRF through unanchored autocomplete endpoint intermediate | This vulnerability report details a Server-Side Request Forgery (SSRF) in code16/sharp, specifically concerning an unanchored autocomplete endpoint. The flaw allows attackers to potentially exploit the autocomplete functionality to make unauthorized requests to internal or external resources. This could lead to compromised data or unauthorized access to systems. The report highlights the critical nature of this security weakness. |
| 2026-11-12 2026 | How an SSRF in a Workflow Tool Led to AWS Credentials and S3 Data Access intermediate Secrets | A Server-Side Request Forgery (SSRF) vulnerability was discovered in a workflow tool, allowing attackers to access sensitive AWS credentials and S3 data. The SSRF flaw enabled the attacker to manipulate requests made by the workflow tool, effectively tricking it into revealing its internal AWS credentials. With these credentials, the attacker gained unauthorized access to Amazon S3 buckets, potentially exposing or exfiltrating confidential data. This highlights a critical security risk in how workflow tools handle internal API requests and credential management. |
| 2026-11-12 2026 | The $8,000 Kubernetes SSRF Bug Bounty: How One Hidden HTTP Header Revealed an Entire Internal… intermediate Bug Bounty | A security researcher discovered a critical Server-Side Request Forgery (SSRF) vulnerability in Kubernetes, earning an $8,000 bounty. The flaw was triggered by a seemingly innocuous, hidden HTTP header. Exploiting this header allowed the researcher to gain access to sensitive internal infrastructure, potentially exposing significant data. This highlights the importance of scrutinizing even minor components of complex systems for security weaknesses. |
| 2026-11-12 2026 | Gotenberg 8.31.0 before 8.38.0 SSRF news | Library entry detailing CVE-2026-126799 affecting Gotenberg versions 8.31.0 through 8.38.0. This Server-Side Request Forgery (SSRF) vulnerability arises from an improper check in the `IsPublicIP()` function, allowing unauthenticated attackers to bypass private IP restrictions by appending IPv6 zone identifiers like `%25eth0` to URLs. This enables attackers to access internal services and exfiltrate responses within generated PDFs. |
| 2026-11-12 2026 | Apache XML Graphics Batik SSRF news | A Server-Side Request Forgery (SSRF) vulnerability has been identified in Apache XML Graphics Batik. This flaw allows an attacker to make unauthorized requests from the server, potentially leading to unauthorized access to internal resources or sensitive data. The vulnerability arises from improper handling of user-supplied URIs within the Batik component. Users are advised to update to the latest version of Apache XML Graphics Batik to mitigate this security risk. |
| 2026-11-12 2026 | Adobe Connect SSRF (critical) news | Adobe Connect is vulnerable to a critical Server-Side Request Forgery (SSRF) flaw. This vulnerability allows an attacker to trick the application into making arbitrary requests on their behalf to internal or external resources. Exploiting this could lead to unauthorized access to sensitive data and potentially further compromise the network. The severity of this SSRF vulnerability warrants immediate attention and patching by users of Adobe Connect. |
| 2026-11-11 2026 | Novu SSRF via conditions filter webhook bypasses validateUrlSsrf() news 4 min read | Writeup on CVE-2026-125114 detailing how Novu's conditions filter webhook bypasses the `validateUrlSsrf()` protection. This vulnerability, stemming from unvalidated `webhookUrl` usage with `axios.post()`, allows for full-read SSRF, potentially exposing cloud metadata like AWS IMDSv1 credentials. The fix involves incorporating `validateUrlSsrf()` before webhook requests, as implemented in Novu version 3.15.0. |
| 2026-11-11 2026 | N-able N-central SSRF (critical) news | A critical Server-Side Request Forgery (SSRF) vulnerability has been identified in N-able N-central. This flaw allows an attacker to force the server to make requests to arbitrary internal or external resources. Exploiting this vulnerability could lead to unauthorized access to sensitive data, network reconnaissance, or even further system compromise. Organizations using N-able N-central should prioritize patching or mitigating this security risk to protect their environments. |
| 2026-11-11 2026 | Adobe Connect SSRF (critical) news | Adobe Connect is vulnerable to a critical Server-Side Request Forgery (SSRF) vulnerability. This flaw allows unauthenticated attackers to perform arbitrary SSRF attacks, potentially leading to unauthorized access to internal systems and data. The vulnerability is severe and requires immediate attention from users of Adobe Connect. |
| 2026-11-11 2026 | Azure Managed Grafana unauthenticated SSRF (critical) news | An unauthenticated Server-Side Request Forgery (SSRF) vulnerability has been discovered in Azure Managed Grafana. This critical flaw allows an attacker to make arbitrary requests on behalf of the Grafana instance. The vulnerability stems from a lack of proper validation in how the service handles external data sources. This could lead to unauthorized access to internal resources or sensitive information within the Azure environment. |
| 2026-11-11 2026 | Microsoft 365 Agents SDK SSRF news AI | A server-side request forgery (SSRF) vulnerability has been discovered in the Microsoft 365 Agents SDK. This flaw allows attackers to make unauthorized requests from the server, potentially leading to access to sensitive internal resources or data. The vulnerability was reported, and Microsoft is likely addressing the issue. |
| 2026-11-10 2026 | Replicate Cog through 0.23.1 SSRF intermediate AI Python | Library for detecting and mitigating Server-Side Request Forgery (SSRF) in Replicate Cog versions prior to 0.23.1. The vulnerability allows attackers to submit prediction inputs with attacker-controlled URLs, bypassing host or IP validation and forcing the Cog process to make unauthorized HTTP requests to internal services or sensitive resources. Exploiting HTTP redirects in cog.File and cog.Path predictor inputs can further facilitate these attacks. |
| 2026-11-10 2026 | Omnom through 0.9.0 SSRF intermediate | Omnom versions prior to 0.9.0 are vulnerable to a Server-Side Request Forgery (SSRF) flaw. This vulnerability allows an attacker to control the requests made by the server. If exploited, an attacker could potentially access internal resources or interact with other services on the network that the Omnom server has access to. Users are advised to update to version 0.9.0 or later to mitigate this security risk. |
Frequently Asked Questions
- What is SSRF and why is it critical?
- Server-Side Request Forgery (SSRF) tricks the server into making HTTP requests to unintended destinations. It is critical because it can access internal services, cloud metadata endpoints (169.254.169.254), and private networks that are not reachable from the internet, potentially leading to full cloud account compromise.
- What are common SSRF bypass techniques?
- Common bypasses include DNS rebinding, IP encoding tricks (decimal, octal, hex, IPv6), URL parser inconsistencies between the filter and the HTTP library, open redirect chains, and alternative URL schemes. Attackers also use shortened URLs and domain names that resolve to internal IPs.
- Where does SSRF commonly appear?
- SSRF is frequently found in webhook integrations, URL preview and unfurling features, PDF and screenshot generators, image/file processing pipelines, and any functionality where the application fetches a remote resource based on user-supplied URLs.
Weekly AppSec Digest
Get new resources delivered every Monday.