hammerspoon

Warn

Audited by Gen Agent Trust Hub on Jul 24, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONDATA_EXFILTRATION
Full Analysis
  • [DYNAMIC_EXECUTION]: The core functionality of the skill is executing arbitrary Lua code on the host system via the hs CLI tool. This allows the agent to generate and run logic at runtime.
  • [DATA_EXPOSURE_EXFILTRATION]: The skill provides recipes to read the system clipboard (hs.pasteboard.getContents()) and extract browser tab URLs and titles via AppleScript. These capabilities allow the agent to access potentially sensitive user information.
  • [PERSISTENCE_MECHANISMS]: The skill explicitly guides the agent to create persistent automations by writing Lua scripts to the ~/.hammerspoon/ directory and modifying init.lua. This configuration is automatically loaded and executed whenever the Hammerspoon application starts.
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted data from the clipboard and browser environment. There is a risk that if this data contains malicious instructions and is used to construct further commands without proper sanitization, it could influence the agent's behavior.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Jul 24, 2026, 01:38 AM
Security Audit — agent-trust-hub — hammerspoon