Do Agents Dream of Root Shells? Partial-Credit Evaluation of LLM Agents in Capture the Flag Challenges

Large Language Model LLM agents are increasingly proposed for autonomous cybersecurity tasks, but their capabilities in realistic offensive settings remain poorly understood. We present DeepRed, an open-source benchmark for evaluating LLM-based agents on realistic Capture The Flag CTF challenges ...

aider-chat (0.82.3), aird (0.2.0) +7 more potentially affected by CVE-2026-3219 via pip (25.1.1)

pip PYPI version 25.1.1 is affected by a known vulnerability. The following packages have a transitive dependency on pip and may be impacted: - aider-chat 0.82.3 - aird 0.2.0 - anki-release =25.6.0b2, =0.8.0, =1.0.0, =1.0.0, =0.1.11, =0.1.12 Source cves: CVE-2026-3219 Source advisory:...

SGLang CVE-2026-5760 (CVSS 9.8) Enables RCE via Malicious GGUF Model Files

A critical security vulnerability has been disclosed in SGLang that, if successfully exploited, could result in remote code execution on susceptible systems. The vulnerability, tracked as CVE-2026-5760 , carries a CVSS score of 9.8 out of 10.0. It has been described as a case of command injection...

Arbitrary Code Injection

Overview sglang is a SGLang is a fast serving framework for large language models and vision language models. Affected versions of this package are vulnerable to Arbitrary Code Injection via the reranking endpoint when a model file containing a malicious tokenizer.chattemplate is loaded, due to...

EUVD-2026-23860

SGLang's reranking endpoint /v1/rerank achieves Remote Code Execution RCE when a model file containing a malcious tokenizer.chattemplate is loaded, as the Jinja2 chat templates are rendered using an unsandboxed jinja2.Environment...

Security Bulletin: DevOps Test Performance contains a vulnerability related to use of Spring Framework MVC and WebFlux

Summary Due to use of Spring Framework MVC and WebFlux, DevOps Test Performance and Rational Performance Tester contain a potential stream corruption vulnerability. Vulnerability Details CVEID:CVE-2026-22735 DESCRIPTION: Spring MVC and WebFlux applications are vulnerable to stream corruption when...

Anthropic MCP Design Vulnerability Enables RCE, Threatening AI Supply Chain

Cybersecurity researchers have discovered a critical "by design" weakness in the Model Context Protocol's MCP architecture that could pave the way for remote code execution and have a cascading effect on the artificial intelligence AI supply chain. "This flaw enables Arbitrary Command Execution R...

GHSA-5H65-JX66-J7P5 FastChat has Denial of Service Through Blocking Event Loop in Model Workers (Incomplete Fix for ff66426)

A security vulnerability has been detected in lm-sys fastchat up to 0.2.36. This issue affects the function apigenerate of the component Worker API Endpoint. The manipulation leads to resource consumption. The attack can be initiated remotely. The exploit has been disclosed publicly and may be...

FastChat has Denial of Service Through Blocking Event Loop in Model Workers (Incomplete Fix for ff66426)

A security vulnerability has been detected in lm-sys fastchat up to 0.2.36. This issue affects the function apigenerate of the component Worker API Endpoint. The manipulation leads to resource consumption. The attack can be initiated remotely. The exploit has been disclosed publicly and may be...

Improper Neutralization of Special Elements in Output Used by a Downstream Component ('Injection')

Overview Affected versions of this package are vulnerable to Improper Neutralization of Special Elements in Output Used by a Downstream Component 'Injection' via the installmcpconfig function in the Model Context Protocol Configuration API when processing the X-Forwarded-For argument. An...

CVE-2026-6607

A security vulnerability has been detected in lm-sys fastchat up to 0.2.36. This issue affects the function apigenerate of the component Worker API Endpoint. The manipulation leads to resource consumption. The attack can be initiated remotely. The exploit has been disclosed publicly and may be...

CVE-2026-6607

CVE-2026-6607 affects lm-sys FastChat up to version 0.2.36, specifically the Worker API Endpoint function api_generate. The issue allows remote manipulation leading to resource consumption; CVE details indicate a publicly disclosed exploit and a patch is available (patch id c9e84b89c91d45191dc244...

EUVD-2026-23729

A weakness has been identified in serge-chat serge up to 1.4TB. The impacted element is the function downloadmodel/deletemodel of the file api/src/serge/routers/model.py of the component Model API Endpoint. Executing a manipulation can lead to missing authentication. The attack can be launched...

CVE-2026-6599

The CVE-2026-6599 entry concerns langflow-ai langflow (up to v1.8.3) with a flaw in the Model Context Protocol Configuration API. The affected element is the file src/backend/base/langflow/api/v1/mcp_projects.py, specifically the install_mcp_config function (and mention of get_client_ip). Manipul...

CVE-2026-6599 langflow-ai langflow Model Context Protocol Configuration API mcp_projects.py install_mcp_config injection

A vulnerability was detected in langflow-ai langflow up to 1.8.3. The impacted element is the function getclientip/installmcpconfig of the file src/backend/base/langflow/api/v1/mcpprojects.py of the component Model Context Protocol Configuration API. Performing a manipulation of the argument...

CVE-2026-6599

A vulnerability was detected in langflow-ai langflow up to 1.8.3. The impacted element is the function getclientip/installmcpconfig of the file src/backend/base/langflow/api/v1/mcpprojects.py of the component Model Context Protocol Configuration API. Performing a manipulation of the argument...

CVE-2026-6599 langflow-ai langflow Model Context Protocol Configuration API mcp_projects.py install_mcp_config injection

A vulnerability was detected in langflow-ai langflow up to 1.8.3. The impacted element is the function getclientip/installmcpconfig of the file src/backend/base/langflow/api/v1/mcpprojects.py of the component Model Context Protocol Configuration API. Performing a manipulation of the argument...

CVE-2026-6590

A vulnerability was detected in ComfyUI up to 0.13.0. This impacts the function getmodelpreview of the file app/modelmanager.py of the component Model Preview Endpoint. The manipulation results in path traversal. The attack may be launched remotely. The exploit is now public and may be used. The...

CVE-2026-6588

A weakness has been identified in serge-chat serge up to 1.4TB. The impacted element is the function downloadmodel/deletemodel of the file api/src/serge/routers/model.py of the component Model API Endpoint. Executing a manipulation can lead to missing authentication. The attack can be launched...

CVE-2026-6590 ComfyUI Model Preview Endpoint model_manager.py get_model_preview path traversal

A vulnerability was detected in ComfyUI up to 0.13.0. This impacts the function getmodelpreview of the file app/modelmanager.py of the component Model Preview Endpoint. The manipulation results in path traversal. The attack may be launched remotely. The exploit is now public and may be used. The...