A Server-Side Request Forgery (SSRF) protection bypass vulnerability exists in the MCP Client node of n8n. Authenticated users with workflow creation or editing permissions could exploit this to make the server connect to internal or blocked hosts, bypassing SSRF protections and potentially exposing internal services. The issue is fixed in n8n versions 2.31.5 and 2.32.1. Temporary mitigations include restricting access to trusted users, disabling the MCP Client node, and limiting network egress to internal address ranges.
A vulnerability in n8n versions from 2.27.0 up to but not including 2.29.8 and 2.30.x before 2.30.1 allows member-level users to execute other users' MCP Server Trigger workflows due to missing OAuth authorization checks. This flaw permits a member-level user to register an OAuth client, self-approve consent for another user's workflow, and obtain a valid token, enabling execution of workflows in the owner's context and access to their data. The advisory is a duplicate and has been withdrawn in favor of GHSA-q5xf-xhwf-cwqf. No patch links are provided in this record.
A flaw was found in the Ansible Lightspeed Model Context Protocol (MCP) server. This vulnerability, known as path traversal, allows an attacker to manipulate an AI agent through indirect prompt injection. By doing so, the attacker can cause the server to write files to unauthorized locations on the user's system. This can result in the exposure of sensitive host information and enable the attacker to execute malicious commands, potentially leading to a full system compromise.
n8n before 2.29.8 and 2.30.x before 2.30.1 (affected from 2.27.0, when the OAuth 2.1 consent and token-issuance flow was introduced) does not verify that the authenticated user has access to the workflow referenced as the OAuth resource. On instances with at least one active MCP Server Trigger workflow configured with n8n OAuth2 authentication, a member-level user can register an OAuth client, self-approve consent for another user's workflow, and obtain a valid token. The workflow then runs in the owner's project context with the owner's stored credentials, and the attacker can set tool inputs and read outputs (potentially including data from the owner's connected integrations), breaking user and project isolation.
mcp-webresearch 0.1.7 contains a server-side request forgery vulnerability that allows attackers to access internal network services by supplying loopback, link-local, or cloud metadata addresses to the visit_page tool, which only validates the URL protocol without filtering private or reserved IP ranges. Attackers can steer the LLM-controlled URL argument through prompt injection to navigate the server's Playwright browser to internal endpoints such as cloud instance metadata services, causing the server to return sensitive internal page content including credentials into the model context.
PraisonAI is a multi-agent teams system. Prior to version 4.6.40, the fix for GHSA-9mqq-jqxf-grvw / CVE-2026-44336 is incomplete. The original advisory description named four vulnerable handlers in `mcp_server/adapters/cli_tools.py`. Commit `68cc9427` ("fix(security): harden MCP rules path handling…") added a `_resolve_rule_path()` helper and applied it to `rules.create`, `rules.show`, and `rules.delete`. `workflow.show` was left unchanged. Two adjacent handlers in the same file have the same pattern, `workflow.validate` and `deploy.validate`. Neither was mentioned in the original advisory. Both remained unchanged. The original advisory also identified the dispatcher (`server.py:281-298`) as a root cause. It accepts unvalidated `**kwargs` from `params["arguments"]` with no enforcement against the tool's declared `input_schema`. That code is unchanged prior to version 4.6.40. A single unauthenticated MCP `tools/call` to `praisonai.workflow.show` returns the contents of any file the host user can read: `/etc/passwd`, `~/.ssh/id_rsa`, `~/.aws/credentials`, or any project `.env`. Version 4.6.40 contains an updated fix.
WhatsApp MCP Server is a Model Context Protocol (MCP) server for WhatsApp, enabling Claude to read and send WhatsApp messages. Prior to version 0.2.1, the `whatsapp-bridge` HTTP API listens on `127.0.0.1:8080` without authentication and without Host header validation, and the `/api/send` endpoint accepts an absolute `media_path` parameter without confining it to a safe directory. Combined, these issues allow any local process running as the same user as the bridge to send WhatsApp messages from the paired account without authorization; the same caller to read arbitrary files readable by the user (e.g. SSH private keys, browser session data, source code, dotfiles) and exfiltrate them as WhatsApp document attachments; and/or a remote attacker to trigger the same operations via DNS rebinding from a webpage the user visits, since no Host header validation is performed. In MCP environments, "local caller" extends beyond processes the user explicitly launched — sibling MCP servers, IDE extensions, and tool-triggered flows running in the user's session can act as the effective caller. This issue is fixed in whatsapp-mcp v0.2.1 and corresponding Docker images / release artifacts. Users should upgrade immediately. The fix introduces bearer token authentication on the bridge HTTP API (configured via environment variable, required on all requests, validated with constant-time comparison); host header allow-list validation to prevent DNS rebinding; and confinement of `media_path` to a configured directory, with rejection of absolute paths outside the root and path traversal sequences. This is a breaking change for clients of the bridge API. For users who cannot immediately upgrade: Stop the bridge, or block loopback access to port 8080, when the bridge is not actively in use; avoid running the bridge alongside untrusted MCP servers, browser extensions, or other untrusted local processes; avoid browsing untrusted sites while the bridge is running (DNS rebinding mitigation); and/or run the bridge under a dedicated user account or in a sandbox/container with no access to sensitive files.
Network-AI is a TypeScript/Node.js multi-agent orchestrator. Prior to version 5.4.5, the MCP SSE server defaults to an empty secret (`process.env['NETWORK_AI_MCP_SECRET'] ?? ''` at `bin/mcp-server.ts:89`), which causes `_isAuthorized` (`lib/mcp-transport-sse.ts:254`) to return `true` unconditionally for every request — no `Authorization` header is required. Simultaneously, `_handleRequest` sets `Access-Control-Allow-Origin: *` (`lib/mcp-transport-sse.ts:272`) on every response, so a cross-origin browser fetch can read the result without restriction. An unauthenticated attacker who can lure a user to a malicious web page can invoke all 22 exposed MCP tools — including `config_set`, `agent_spawn`, and `blackboard_write` — against a default-configured localhost server. Version 5.4.5 patches the issue.
CVE-2026-15415 is a medium severity vulnerability in AWS HealthOmics MCP Server versions before 0.0.36. It involves a relative path traversal flaw in the linting tools that could allow an attacker who can influence the MCP agent to write files to arbitrary locations outside the intended workflow bundle directory. This could lead to unauthorized file writes. AWS has released version 0.0.36 to remediate this issue.
IBM Langflow OSS 1.0.0 through 1.10.0 Langflow versions up to 1.9.2 (commit 94981c443d4918517b9e8163d70fc598dc33a32d) contain a code injection vulnerability in the Policies component's ToolGuard integration that bypasses the allow_custom_components=false security control. The vulnerability exists because the validation mechanism only checks the main component source code in node_template["code"]["value"] but fails to validate dynamic CodeInput fields that store generated ToolGuard Python files. Attackers can embed malicious Python code in these unvalidated dynamic fields, which are persisted in Flow.data and later executed server-side when a guarded tool is invoked through the ToolGuard runtime. This allows authenticated users with flow creation privileges to achieve arbitrary Python code execution on the backend despite custom component restrictions. The vulnerability can be escalated through cross-tenant flow manipulation via the agentic MCP update_flow_component_field tool, which accepts attacker-controlled user_id parameters, enabling attackers to inject malicious code into victim users' flows. When combined with publicly accessible flows and specific misconfigurations (AUTO_LOGIN=true, NEW_USER_IS_ACTIVE=true), the attack can be conducted with reduced authentication requirements.
Agentic-Flow is an AI agent orchestration platform. Prior to 2.0.14, agentic-flow MCP server tools in src/mcp/standalone-stdio.ts, src/mcp/fastmcp/servers/claude-flow-sdk.ts, src/mcp/fastmcp/servers/stdio-full.ts, src/mcp/fastmcp/servers/http-streaming-updated.ts, src/mcp/fastmcp/servers/http-sse.ts, src/mcp/fastmcp/servers/poc-stdio.ts, src/mcp/fastmcp/tools/agent/{execute,list,parallel}.ts, src/mcp/fastmcp/tools/swarm/orchestrate.ts, and src/mcp/fastmcp/tools/hooks/pretrain.ts interpolated attacker-influenceable tool parameters such as agent, task, name, language, and agentdb directly into shell command strings passed to execSync(), allowing arbitrary OS command execution with the privileges of the MCP server user. This issue is fixed in version 2.0.14.
---
_-= Per source details. Do not edit below this line.=-_
## Source: amazon-inspector (6693129f8b6f8b6c2d52722ca7746d6ef3dd792476297fa3583d3956548eb2b4)
The package ships heavily obfuscated runtime files (dist/index.js, dist/launcher.js) built with javascript-obfuscator. The MCP server registered as the package's main entry exposes an `install_mcp` tool that fetches JSON from the hardcoded portal `prompt-injection-tool-portal.vercel.app` and passes its `install_command` field directly to `child_process.exec()` on the installer's host, giving the portal operator arbitrary shell execution with the installer's privileges. The launcher (invoked by every `salesbot1..10` bin entry) calls `fetchBundledMcps` against the same portal and, for each returned entry, spawns `npx -y <entry.package>` while injecting locally decrypted vault secrets into the child's env — the portal can name any npm package, including a freshly published attacker-owned one, and it will be downloaded and executed with those secrets on hand. The launcher also unconditionally spawns a detached `npm i -g @funny-booth/agent-core@latest` on every run, silently self-updating to whatever the publisher pushes next. The launcher additionally spawns the local `claude` CLI with a portal-supplied greeting/knowledge string prepended as the initial user prompt and a portal-controlled MCP config, providing a prompt-injection channel into the user's Claude agent session. The repository URL in package.json is `github.com/yutamatsuura/prompt-injection-tool` and the deliberate obfuscation of the portal endpoints, exec sink, and auto-update spawn is consistent with intentional concealment of these remote-execution channels.
dbt-mcp is a Model Context Protocol server for interacting with dbt. Prior to 1.17.1, DbtMCP.call_tool() in src/dbt_mcp/mcp/server.py logged the raw arguments dictionary at INFO level before each tool call and at ERROR level on exceptions, and configure_file_logging() wrote those records to dbt-mcp.log when DBT_MCP_SERVER_FILE_LOGGING=true, preserving sensitive sql_query, vars, and node_selection values in plaintext without automatic rotation or deletion. This issue is fixed in version 1.17.1.
dbt-mcp is a Model Context Protocol server for interacting with dbt. Prior to 1.17.1, DefaultUsageTracker.emit_tool_called_event() in src/dbt_mcp/tracking/tracking.py serialized every MCP tool call's complete arguments dictionary and sent it through dbtlabs_vortex.producer.log_proto without redaction, including sql_query from show, vars from run, build, and test, and node_selection from compile, while usage_tracking_enabled in settings.py enabled telemetry by default unless DBT_SEND_ANONYMOUS_USAGE_STATS=false or DO_NOT_TRACK=1 was set. This issue is fixed in version 1.17.1.
dbt-mcp is a Model Context Protocol server for interacting with dbt. Prior to 1.17.1, _run_dbt_command() in src/dbt_mcp/dbt_cli/tools.py appended unsanitized node_selection and resource_type values to the dbt subprocess argument list, allowing an MCP client to inject dbt global flags such as --profiles-dir, --project-dir, and --target into subprocess.Popen even though shell=False prevents shell metacharacter injection. This issue is fixed in version 1.17.1.
The Apify MCP server enables AI agents to extract data from websites using ready-made scrapers, crawlers, and automation tools available on the Apify Store. Prior to 0.9.21, the fetch-apify-docs tool in src/tools/common/fetch_apify_docs.ts validates allowlisted documentation domains with String.startsWith() rather than URL hostname comparison, allowing attacker-controlled URLs such as `https://docs.apify.com.evil.com/` and `https://docs.apify.com@evil.com/` to pass the ALLOWED_DOC_DOMAINS check and return arbitrary fetched content to the LLM. This issue is fixed in version 0.9.21.
ToolHive is a utility designed to simplify the deployment and management of Model Context Protocol (MCP) servers. Prior to 0.31.0, remote.Handler.Authenticate in pkg/auth/remote/handler.go invokes discovery.DetectAuthenticationFromServer in pkg/auth/discovery/discovery.go, whose host-side HTTP clients trust remote-server-controlled authentication discovery destinations, follow redirects without host or scheme restrictions, and do not consistently block private addresses. A malicious or compromised remote MCP server can place a resource_metadata URL in WWW-Authenticate for ParseWWWAuthenticate to extract, after which FetchResourceMetadata, OIDC issuer discovery, and well-known discovery can issue GET requests to link-local, RFC1918, or other internal services outside the server's container. The user connects to a server that the user intends to use, but the server controls the internal destination; this path does not depend on the separate NAT64 guard omission because the affected clients did not invoke IsPrivateIP, and it differs from the DCR resolver path that already refused redirects. The demonstrated primitive reaches internal-only HTTP services and reachability or error oracles, and can retrieve AWS metadata credentials where IMDSv1 accepts an unauthenticated GET, while IMDSv2 and GCP metadata prerequisites are not satisfied by the demonstrated request. This vulnerability is fixed in 0.31.0.
@andrea9293/mcp-documentation-server version 1.13.0 exposes its Web UI and document-management API on all network interfaces by default without requiring authentication. This allows any client on the same LAN or network segment to perform administrative actions such as listing, adding, reading, searching, and deleting documents without credentials. The issue arises because the server binds to 0.0.0.0 instead of localhost only, and no authentication is enforced on API endpoints.
### Summary
In affected versions, the deprecated WebSocket server transport (`mcp.server.websocket.websocket_server`) accepted the WebSocket handshake without applying any `Host` or `Origin` header validation. The `TransportSecuritySettings` mechanism that the SSE and Streamable HTTP transports use for this purpose was not wired into the WebSocket transport, so there was no SDK-level way to restrict which origins could connect.
### Am I affected?
Only if a developer's application server exposes `mcp.server.websocket.websocket_server`. This transport has never been part of the MCP specification, is marked deprecated, and is not reachable through `FastMCP` — a developer must have wired it into an ASGI application themselves. Servers using stdio, SSE, or Streamable HTTP are not affected by this advisory.
### Details
`websocket_server()` constructed a Starlette `WebSocket` and called `accept(subprotocol="mcp")` immediately, with no inspection of the connection's headers. By contrast, `SseServerTransport` and `StreamableHTTPServerTransport` accept an optional `security_settings: TransportSecuritySettings` and run `TransportSecurityMiddleware.validate_request()` against the incoming `Host` and `Origin` headers before establishing a session. Because browsers attach an `Origin` header to cross-origin WebSocket upgrade requests but do not enforce a same-origin policy on the response, a web page served from any origin could open a WebSocket to a reachable MCP server on this transport, complete the `initialize` handshake, and issue JSON-RPC requests on the resulting session.
### Impact
A user who runs an MCP server on this transport bound to localhost or a LAN address, without a separate authentication or origin gate in front of it, and visits a malicious web page, can have that page enumerate and invoke the server's tools and read its resources. The consequences depend entirely on what the server exposes. The transport itself requires no token or prior session. Some browsers prompt before allowing a public page to open a connection to a local-network address, which adds a user-interaction step but is not a substitute for server-side validation.
### Mitigation
Upgrade to version 1.28.1 or later, in which `websocket_server()` accepts the same optional `security_settings: TransportSecuritySettings` argument as the other HTTP-based transports and validates the `Host` and `Origin` headers before accepting the handshake; a request that fails validation is rejected with HTTP 403 and `ValueError("Request validation failed")` is raised to the caller. As with the other transports the parameter defaults to `None`, which leaves validation disabled, so upgrading alone does not change behaviour: pass a `TransportSecuritySettings` with `enable_dns_rebinding_protection=True` and appropriate `allowed_hosts` / `allowed_origins` to receive the protection. The recommended path remains to migrate off this deprecated transport to Streamable HTTP, where `FastMCP` enables this protection automatically for localhost binds. The WebSocket transport has been removed entirely in v2.
## Impact
In multi-tenant HTTP deployments — where a single n8n-mcp server serves several tenants — the locally stored workflow version history (the automatic backups taken before workflow updates) was not isolated per tenant. An authenticated tenant could read workflow version snapshots belonging to other tenants, and could delete or destroy other tenants' stored backups.
A stored snapshot includes full node definitions, so the exposed data can contain credential references and authorization headers configured on nodes. This is therefore a confidentiality issue in addition to an integrity/availability one.
## Affected configurations
- HTTP mode with multi-tenancy enabled (`ENABLE_MULTI_TENANT=true`), where multiple tenants are served by a single shared instance and database.
Not affected:
- stdio / single-user deployments (e.g. Claude Desktop).
- Single-tenant HTTP deployments (one tenant per instance and database).
## Affected versions
`<= 2.56.0`
## Patched version
`2.56.1`. The stored version history is now isolated per instance, so a tenant can only access its own backups. Upgrading runs a one-time migration that isolates existing history and clears previously stored, un-scoped backups (these are auto-created, short-retention backups).
## Workarounds
If users cannot upgrade immediately:
- Disable the workflow version tool by setting `DISABLED_TOOLS=n8n_workflow_versions` in the server environment (for example, in your Docker `.env`). This removes the affected tool from the deployment for all tenants; automatic backups are unaffected, but the cross-tenant access path is closed.
- Alternatively, do not run in multi-tenant mode — serve each tenant from a separate instance with its own database, so no local store is shared between tenants.
- Restrict network access to the HTTP endpoint to trusted operators.
stdio and single-tenant HTTP deployments are not affected.
## Credit
Reported by Francisco Rosales (@0xmagic0) and coordinated by Ax Sharma (@axsharma) of Manifold Security.