automation-mcp

CRITICAL E005: Suspicious download URL detected in skill instructions.

• Suspicious download URL detected (high risk: 0.85). The GitHub URL is a personal/unverified repository that distributes code which grants full desktop automation and should not be trusted or executed without review (high-risk); the bun.sh link is the official Bun installer and is not flagged.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). The skill explicitly grants a remote assistant full desktop control (mouse/keyboard/screen capture) and instructs installing/running a locally-hosted MCP server, which enables easy credential capture, screen/data exfiltration, and remote control—high abuse potential and functions like a backdoor.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The SKILL.md directs executing fetched remote code during setup (e.g., "curl -fsSL https://bun.sh/install | bash" and "git clone https://github.com/ashwwwin/automation-mcp.git" then running the cloned code), so these URLs fetch and install/execute code required for the agent to run.

MEDIUM W013: Attempt to modify system services in skill instructions.

• Attempt to modify system services in skill instructions detected (high risk: 1.00). This skill explicitly provisions an MCP automation server that gives the agent full mouse/keyboard control, screenshots, and window management and instructs the user to grant macOS Accessibility/Screen Recording permissions and install/run code (via curl|bash and bun), which enables the agent to change the machine's state and perform privileged-like interactive actions.