CVE-2026-27825

CVE-2026-27825 — Path Traversal in mcp-atlassian via confluenceuploadattachment Severity : Critical CVSS 9.3 CWE : CWE-22 — Path Traversal Affected : sooperset/mcp-atlassian = 0.17.0 read-side twin of GHSA-xjgw-4wvw-rgm4 Advisory : GHSA NVD : https://nvd.nist.gov/vuln/detail/CVE-2026-27825 TL;DR...

CVE-2026-59822

A flaw was found in LiteLLM, a proxy server designed for Large Language Model LLM APIs. An unauthenticated attacker could exploit a vulnerability in the MCP Streamable HTTP endpoint. By using a fabricated Authorization header, the attacker could bypass the standard key validation, gaining...

CVE-2026-53710

creationtimestamp| type| source ---|---|--- 2026-09-02 16:35:09+00:00| published-proof-of-concept| https://github.com/IBM/mcp-context-forge/security/advisories/GHSA-xm98-3vcf-fph7 2026-09-10 12:30:07+00:00| seen| https://bsky.app/profile/stackflag.bsky.social/post/3mv62kfu7xv2r 2026-09-15...

EUVD-2026-58725

CKAN MCP Server: MQA server allowlist bypass via unanchored regex isValidMqaServer...

SUSE CVE-2026-81092

mcp-go accepted requests on its HTTP transports without checking the Host header. StreamableHTTPServer.ServeHTTP in server/streamablehttp.go and SSEServer.ServeHTTP in server/sse.go served any request arriving over a loopback connection regardless of the host it named, and the SSE transport's...

Security Bulletin: IBM ContextForge MCP Gateway is affected by server-side request forgery (DNS rebinding) via the A2A agent invocation endpoint

Summary IBM ContextForge MCP Gateway contains a server-side request forgery SSRF vulnerability in its Agent-to-Agent A2A invocation endpoint POST /a2a/agentname/invoke. Unlike every other outbound egress path in the gateway, the A2A invocation path does not pin the DNS-resolved IP address at...

Security Bulletin: IBM ContextForge MCP Gateway is affected by credential disclosure and privilege escalation via jq filter execution

Summary IBM ContextForge MCP Gateway allows authenticated users with developer privileges to supply a jq program in the tool jsonpathfilter field that executes unrestricted built-in such as $ENV inside the server process, exposing environment variables including cryptographic keys and database...

github.com/mark3labs/mcp-go: mcp-go: DNS Rebinding vulnerability due to missing Host header validation

A flaw was found in mcp-go. This component failed to validate the Host header in HTTP requests, specifically within its StreamableHTTPServer and SSEServer transports. A remote attacker could exploit this by using a crafted web page to perform a DNS rebinding attack. This allows the attacker to...

github.com/mark3labs/mcp-go: mcp-go: DNS Rebinding vulnerability due to missing Host header validation

A flaw was found in mcp-go. This component failed to validate the Host header in HTTP requests, specifically within its StreamableHTTPServer and SSEServer transports. A remote attacker could exploit this by using a crafted web page to perform a DNS rebinding attack. This allows the attacker to...

EUVD-2026-69713

A vulnerability was found in NousResearch hermes-agent up to 0.18.2. This vulnerability affects the function listtools of the file tools/mcptool.py of the component MCP Tool. Performing a manipulation results in uncontrolled memory allocation. It is possible to initiate the attack remotely. The...

BerriAI LiteLLM Improper Authentication Vulnerability

BerriAI LiteLLM contains an improper authentication vulnerability in the MCP Streamable HTTP endpoint that could allow an unauthenticated attacker to establish an authenticated MCP session using an arbitrary Bearer token...

CVE-2026-84289

A vulnerability was found in NousResearch hermes-agent up to 0.18.2. This vulnerability affects the function listtools of the file tools/mcptool.py of the component MCP Tool. Performing a manipulation results in uncontrolled memory allocation. It is possible to initiate the attack remotely. The...

CVE-2026-84289

A vulnerability in NousResearch hermes-agent up to version 0.18.2 affects the list_tools function in tools/mcp_tool.py (MCP Tool component). The root cause is uncontrolled memory allocation , which can lead to a resource amplification denial-of-service when manipulated remotely. The CVSS 4.0 scor...

CVE-2026-84289 NousResearch hermes-agent MCP Tool mcp_tool.py list_tools memory allocation

A vulnerability was found in NousResearch hermes-agent up to 0.18.2. This vulnerability affects the function listtools of the file tools/mcptool.py of the component MCP Tool. Performing a manipulation results in uncontrolled memory allocation. It is possible to initiate the attack remotely. The...

CVE-2026-81846

An authorization bypass in the runZero Platform MCP service has been resolved in version 5.1.260826.0. This issue is an instance of CWE-639: Authorization Bypass Through User-Controlled Key and has an estimated CVSS score of CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:C/C:L/I:N/A:N 3.5 Low...

CVE-2026-81846 runZero MCP 'Findings summaries' Data Leak

An authorization bypass in the runZero Platform MCP service has been resolved in version 5.1.260826.0. This issue is an instance of CWE-639: Authorization Bypass Through User-Controlled Key and has an estimated CVSS score of CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:C/C:L/I:N/A:N 3.5 Low...

CVE-2026-81846

CVE-2026-81846 is an authorization bypass in the runZero Platform MCP service , classified as CWE-639 (Authorization Bypass Through User-Controlled Key). The vulnerability allows a low-privilege network attacker to bypass authorization and access findings summaries data, resulting in a limited co...

EUVD-2026-69554

An authorization bypass in the runZero Platform MCP service has been resolved in version 5.1.260826.0. This issue is an instance of CWE-639: Authorization Bypass Through User-Controlled Key and has an estimated CVSS score of CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:C/C:L/I:N/A:N 3.5 Low...

CVE-2026-19591

OpenAI Codex CLI (Windows, macOS, Linux) and Codex Desktop (Windows, macOS) contain an approval-bypass flaw in their command-safety parser. The parser misinterprets PowerShell's stop-parsing token (--%), classifying certain commands as safe when they are not. An attacker who controls a repository...

EUVD-2026-69394

OpenAI Codex CLI for Windows, macOS, and Linux and Codex Desktop for Windows and macOS misclassified certain PowerShell commands as safe because their command-safety parser interpreted PowerShell's stop-parsing token --% differently than PowerShell itself. If a user opens an attacker-prepared...