Back to skill

Security audit

MCPGlue — MCP桥接框架

Security checks for vulnerabilities and agentic risk

Overview

This skill is an MCP bridge, but it can launch user-specified local server commands and its documentation overstates capabilities and gives inconsistent install instructions.

Review before installing. Use this only with MCP servers and commands you already trust, and understand that the chosen server process can act with your local user privileges. The publisher should align the docs with the Python implementation, pin dependencies, and add clear warnings or allowlists before this is treated as a low-friction install.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (2)

T09 · Insecure Skill Coding Practices

Error
Location
scripts/mcp_client.py:93
Finding

Unrestricted Local Executable Launch Through User-Controlled Server Command

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Warning
Location
SKILL.md:16
Finding

Incorrect and Unpinned MCP Dependency Installation

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (5)

Tp4

High
Category
MCP Tool Poisoning
Confidence
91% confidence
Finding

The documented behavior materially overstates the implemented controls and capabilities, claiming a broader MCP bridge with SSE, resource access, streaming, dynamic registration, and a specific SDK, while the described implementation appears to be a much simpler stdio client. This mismatch is dangerous because operators may trust the skill to provide security properties or protocol handling it does not actually implement, leading to unsafe deployment assumptions and accidental exposure of external services.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
92% confidence
Finding

The skill exposes MCP connectivity to external services and dynamic tool invocation but does not declare any explicit tool scope, permissions, or allowed-tools restrictions. In an agent environment, this can lead to overbroad access and unsafe execution of external capabilities because consumers cannot tell what systems the skill is permitted to reach or invoke.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
86% confidence
Finding

The skill description explicitly states it can '安全地读写文件' (safely read and write files), which is a user-data-affecting operation. In this markdown file there is no accompanying warning, limitation, or disclosure about possible modification of local files or the need for user caution.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
90% confidence
Finding

The client accepts a user-supplied --server command and executes it through the MCP stdio transport, which will start an external process. Although this appears to be intended functionality for an MCP bridge, the script provides no explicit safety warning, allowlist, or trust boundary guidance, so users may unknowingly launch arbitrary local binaries or scripts with their own privileges.

Content

No source excerpt is available for this finding.

Rp1

Low
Category
MCP Rug Pull
Confidence
60% confidence
Finding

pip install without ==version installs the latest release, which could include malicious changes.

Content

No source excerpt is available for this finding.