Back to skill

Security audit

Cmux

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed local cmux control helper, but it can type into or close active terminal workspaces if an agent uses those commands.

Install this only if you want your agent to control a local cmux session. Prefer read-only status/listing commands unless you intentionally want the agent to type into terminals, press keys, split panes, or close workspaces, and verify the target workspace or surface before mutating actions.

Vulnerability Patterns
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
88% confidence
Finding

The skill exposes environment-dependent behavior via CMUX_SOCKET_PATH and operational control over a local Unix socket, but it declares no explicit tool scope or permission boundaries. In an agent setting, lack of declared scope can cause the skill to be invoked more broadly than intended, increasing the chance of unauthorized interaction with a live terminal multiplexer session.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill advertises commands that can send arbitrary text to terminals and close workspaces, which are effectively command-injection and destructive-control primitives when used against an active cmux session. Without prominent warnings, confirmation requirements, or safe usage guidance, an agent may issue unintended commands that execute in terminal panes, terminate workflows, or alter user state.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.