vp-macos-clean-uninstall

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes various macOS system utilities and package managers through the provided detect-app.sh and scan-residue.sh scripts. These include find, brew, pkgutil, mdls, defaults, npm, pip3, and cargo. All executions are intended for read-only information gathering to assist the agent in identifying application footprints.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes data from the local file system and package manager outputs, which are used to generate a removal plan. To mitigate risks such as search term ambiguity or unintended matches, the scripts implement input validation checks that reject empty strings and common system keywords (e.g., 'mail', 'code', 'sync'). The overall workflow emphasizes user oversight by requiring approval before any destructive commands are run based on the gathered data.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 07:48 AM
Security Audit — agent-trust-hub — vp-macos-clean-uninstall