CVE-2026-54140 · AI agents & MCP
NO FIX YET · SEE WHAT TO DO
OpenHuman, an open-source agent harness, passes names and descriptions from configured external MCP servers straight into the model's context without filtering. A malicious or compromised MCP server can hide instructions in those descriptions that make the agent call other tools, disclose data, or take unauthorized actions.
first public Oct 9, 2026 · 1 source
Check it with your coding agent
It checks your repository and tells you if you’re affected, with the fix.
Only installations that configure an external MCP server are exposed, and the risk depends on that server being attacker-controlled or compromised. Teams that run OpenHuman with no external MCP servers configured can skip it.
Not affected if no-external-mcp-servers, trusted-mcp-servers-only
Who cares ai-agents, desktop, mcp, openhuman
0.57.130.57.13 or laterOpenHuman is an open-source agent harness available as a desktop application and Rust library. Before 0.57.13, OpenHuman places names, descriptions, titles, and parameter definitions returned by configured external MCP servers into the agent model context without neutralizing instructions or bounding the descriptive fields. The McpRemoteTool records produced by list_tools preserve this remote content, while tool filtering considers the tool name and the prompt-injection detector screens user messages rather than MCP definitions. A malicious or compromised MCP server can therefore return a poisoned description that is evaluated on every agent turn and may direct the model to invoke other available tools, disclose accessible data, or perform unauthorized actions without a separate malicious user message. The issue affects users who configure an attacker-controlled or compromised MCP server. This issue is fixed in version 0.57.13.
Are you affected?
Tell InfraStack what you run. It checks this one, and every next one, against your stack.
Ask from here. InfraStack checks the affected versions against your stack and answers in your inbox.
InfraStack checks the affected versions against your stack and answers in Slack, with the evidence and the fix.
NVD data is provided by NIST and is public domain. Posts quoted belong to their authors.