Clipboard Content Factory

• Outcome: Pass
• Latest audit: Jul 7, 2026 · 4:34 AM UTC
• Version: 1.0.1

Overview

This skill is a disclosed clipboard-to-local-drafts automation, but users should be careful because watch mode can save whatever text is copied.

Use this only if you intentionally want clipboard monitoring and local draft generation. Prefer one-shot mode for sensitive environments, avoid running watch mode while copying passwords, tokens, private messages, or business data, and periodically delete the generated Desktop output folder if it may contain sensitive snippets.

SkillSpector

Vulnerability Patterns

• Output Handling: Unvalidated Output Injection, Cross-Context Output, Unbounded Output
• Trigger Abuse: Overly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
• Behavioral AST: exec() Call, eval() Call, Dynamic Import
• MCP Least Privilege: Underdeclared Capability, Wildcard Permission, Missing Permission Declaration
• MCP Tool Poisoning: Hidden Instructions, Unicode Deception, Parameter Description Injection

Findings (8)

subprocess module call

Medium

• Category: Dangerous Code Execution

Content

else:
        import subprocess
        try:
            subprocess.run(
                ["powershell", "-Command", f"Set-Clipboard -Value '{text}'"],
                capture_output=True, timeout=3
            )

• Confidence: 98% confidence
• Finding: The code interpolates raw clipboard text into a PowerShell `-Command` string: `Set-Clipboard -Value '{text}'`. If clipboard content contains quotes or PowerShell metacharacters, it can break out of the quoted string and execute arbitrary PowerShell commands, turning untrusted clipboard data into code execution.

Lp3

Medium

• Category: MCP Least Privilege
• Confidence: 85% confidence
• Finding: The skill advertises clipboard monitoring and file generation but does not declare permissions for shell execution or file writes, despite clearly instructing the use of Python scripts that read clipboard data and save outputs to the desktop. This creates a transparency and consent problem: users and platforms cannot accurately assess what system capabilities the skill will exercise before it runs.

Tp4

High

• Category: MCP Tool Poisoning
• Confidence: 93% confidence
• Finding: The documented behavior overstates and obscures what the skill actually does: it continuously monitors the clipboard, may fetch remote content from copied links, can write generated content back to the clipboard, and stores files locally, yet it frames this as simple 'copy=publish' automation. This mismatch is dangerous because users may unintentionally expose sensitive clipboard contents or trigger local/network side effects they did not meaningfully consent to.

Vague Triggers

Medium

• Confidence: 82% confidence
• Finding: The trigger phrases are broad enough to activate on ordinary requests about clipboard monitoring or automatic content processing, which raises the chance of accidental invocation. Because this skill reads clipboard contents and can perform automated transformations and writes, unintended triggering can cause privacy leakage or unwanted file creation from unrelated copied material.

Missing User Warnings

High

• Confidence: 95% confidence
• Finding: The description does not clearly warn users that the skill can continuously monitor clipboard contents, retrieve linked content over the network, and automatically write generated files to the desktop. Clipboard contents often contain sensitive material, so silent or poorly disclosed monitoring materially increases privacy and data-handling risk in this context.

Missing User Warnings

Medium

• Confidence: 93% confidence
• Finding: Clipboard contents may contain passwords, tokens, private messages, or proprietary text, and this code automatically persists that data to desktop files without confirmation or privacy warning. In a continuously used desktop environment, this creates a realistic confidentiality risk through unintended data retention and exposure to other local users or backup/sync systems.

Missing User Warnings

Medium

• Confidence: 90% confidence
• Finding: Continuous clipboard monitoring is inherently sensitive because users routinely copy secrets and personal data. Running a background watcher without a clear runtime warning, consent flow, or filtering increases the chance of silently collecting and transforming confidential information.

Unvalidated Output Injection

High

• Category: Output Handling

Content

else:
        import subprocess
        try:
            subprocess.run(
                ["powershell", "-Command", f"Set-Clipboard -Value '{text}'"],
                capture_output=True, timeout=3
            )

• Confidence: 99% confidence
• Finding: This is a true injection issue because untrusted clipboard text is embedded directly into a PowerShell command string. A crafted clipboard value containing a closing quote and additional PowerShell statements can cause arbitrary command execution under the current user context.

Static analysis

No suspicious patterns detected.