Quantifying Frontier LLM Capabilities for Container Sandbox Escape

Large language models LLMs increasingly act as autonomous agents, using tools to execute code, read and write files, and access networks, creating novel security risks. To mitigate these risks, agents are commonly deployed and evaluated in isolated "sandbox" environments, often implemented using...

A Systematic Study of LLM-Based Architectures for Automated Patching

Large language models LLMs have shown promise for automated patching, but their effectiveness depends strongly on how they are integrated into patching systems. While prior work explores prompting strategies and individual agent designs, the field lacks a systematic comparison of patching...

GHSA-9H8M-3FM2-QJRQ vulnerabilities

Vulnerabilities for packages: datadog-agent, istio-fips, vitess, gitlab-operator, calico, node-feature-discovery-fips, longhorn-manager-fips, onepassword-operator, docker-cli-buildx-fips, tfsec, gitlab-workhorse-ce-fips, kaniko, conftest-fips, task-fips, cass-operator-fips-no-pvc-delete,...

ClawJacked Flaw Lets Malicious Sites Hijack Local OpenClaw AI Agents via WebSocket

OpenClaw has fixed a high-severity security issue that, if successfully exploited, could have allowed a malicious website to connect to a locally running artificial intelligence AI agent and take over control. "Our vulnerability lives in the core system itself – no plugins, no marketplace, no...

CVE-2026-27896 vulnerabilities

Vulnerabilities for packages: osv-scanner, jaeger, datadog-agent, opencost, gptscript, github-mcp-server, ferretdb...

GHSA-WVJ2-96WP-FQ3F vulnerabilities

Vulnerabilities for packages: osv-scanner, jaeger, datadog-agent, opencost, gptscript, github-mcp-server, ferretdb...

CVE-2026-27734

Beszel is a server monitoring platform. Before v0.18.2, the hub’s authenticated API endpoints GET /api/beszel/containers/logs and GET /api/beszel/containers/info pass the user-supplied container param to the agent without validation. The agent builds Docker Engine API URLs using fmt.Sprintf with ...

GHSA-WVJ2-96WP-FQ3F vulnerabilities

Vulnerabilities for packages: flux-operator, datadog-agent, opencost-fips, jaeger, ferretdb, osv-scanner, jaeger-fips, gitlab-workhorse-ce, datadog-agent-fips, opencost, gptscript, github-mcp-server, gitlab-workhorse-ce-fips, flux-operator-fips...

CVE-2026-27896 vulnerabilities

Vulnerabilities for packages: flux-operator, datadog-agent, opencost-fips, jaeger, ferretdb, osv-scanner, jaeger-fips, gitlab-workhorse-ce, datadog-agent-fips, opencost, gptscript, github-mcp-server, gitlab-workhorse-ce-fips, flux-operator-fips...

GHSA-3645-FXCV-HQR4 Langflow has Remote Code Execution in CSV Agent

Summary The CSV Agent node in Langflow hardcodes allowdangerouscode=True, which automatically exposes LangChain’s Python REPL tool pythonreplast. As a result, an attacker can execute arbitrary Python and OS commands on the server via prompt injection, leading to full Remote Code Execution RCE...

Langflow has Remote Code Execution in CSV Agent

Summary The CSV Agent node in Langflow hardcodes allowdangerouscode=True, which automatically exposes LangChain’s Python REPL tool pythonreplast. As a result, an attacker can execute arbitrary Python and OS commands on the server via prompt injection, leading to full Remote Code Execution RCE...

CVE-2026-27967

Zed, a code editor, has a symlink escape vulnerability in versions prior to 0.225.9 in Agent file tools readfile, editfile. It allows reading and writing files outside the project directory when a project contains symbolic links pointing to external paths. This bypasses the intended workspace...

CVE-2026-27966

Langflow is a tool for building and deploying AI-powered agents and workflows. Prior to version 1.8.0, the CSV Agent node in Langflow hardcodes allowdangerouscode=True, which automatically exposes LangChain’s Python REPL tool pythonreplast. As a result, an attacker can execute arbitrary Python an...

Command Injection

Overview Affected versions of this package are vulnerable to Command Injection in the osctrl-admin environment configuration. An attacker can execute arbitrary shell commands on every endpoint that enrolls using a compromised environment by injecting commands into the hostname parameter, which ar...

Jailbreak Foundry: From Papers to Runnable Attacks for Reproducible Benchmarking

Jailbreak techniques for large language models LLMs evolve faster than benchmarks, making robustness estimates stale and difficult to compare across papers due to drift in datasets, harnesses, and judging protocols. We introduce JAILBREAK FOUNDRY JBF, a system that addresses this gap via a...

The vulnerability of the Data Collection Agent service in the Cisco Catalyst SD-WAN Manager centralized network management system allows a malicious actor to gain unauthorized access to protected information and enhance their privileges.

The vulnerability of the Data Collection Agent service in the Cisco Catalyst SD-WAN Manager centralized network management system is related to the storage of passwords in a recoverable format. Exploiting this vulnerability could allow an attacker to gain unauthorized access to protected...

The vulnerability of the Data Collection Agent service in the Cisco Catalyst SD-WAN Manager centralized network management system allows a perpetrator to gain unauthorized access to protected information.

The vulnerability of the Data Collection Agent service in the Cisco Catalyst SD-WAN Manager centralized network management system is related to errors when privileged application programming interfaces APIs are used. Exploiting this vulnerability could allow a malicious actor to gain unauthorized...

CVE-2026-20128

A vulnerability in the Data Collection Agent DCA feature of Cisco Catalyst SD-WAN Manager could allow an unauthenticated, remote attacker to gain DCA user privileges on an affected system. This vulnerability is due to the presence of a credential file for the DCA user on an affected system. An...

CVE-2026-1229 vulnerabilities

Vulnerabilities for packages: datadog-agent, crossplane-provider-aws-sqs, openbao-fips, crossplane-provider-aws-sns, terraform-provider-azurerm-fips, kubevela, ratify-fips, gitlab-rails-ce, omnictl-multiarch-fips, gitlab-operator, gitlab-runner-fips, omni, buildkitd, cilium-cli, sops-fips,...

GHSA-Q9HV-HPM4-HJ6X vulnerabilities

Vulnerabilities for packages: datadog-agent, crossplane-provider-aws-sqs, openbao-fips, crossplane-provider-aws-sns, terraform-provider-azurerm-fips, kubevela, ratify-fips, gitlab-rails-ce, omnictl-multiarch-fips, gitlab-operator, gitlab-runner-fips, omni, buildkitd, cilium-cli, sops-fips,...