mac-maintenance

Pass

Audited by Gen Agent Trust Hub on Aug 21, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill makes extensive use of shell commands to manage system packages, services, and runtimes. The workflow includes both read-only operations (outdated, list, dry-run) and modifying operations (upgrade, cleanup, install).
  • [COMMAND_EXECUTION]: The skill uses sudo specifically for the softwareupdate utility. While this is a high-privilege operation, it is restricted to official macOS system updates and includes explicit instructions to warn the user and seek consent before proceeding.
  • [EXTERNAL_DOWNLOADS]: The skill triggers updates from well-known, official repositories including the Homebrew registry, the Mac App Store, and runtime-specific registries (NPM, Rustup). These operations are standard for the tool's intended purpose.
  • [PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by ingesting data from external tool outputs.
  • Ingestion points: Data enters the agent context through the output of commands like brew outdated, mas outdated, npm outdated, and brew services list in SKILL.md.
  • Boundary markers: Instructions mandate that the agent must show lists to the user and obtain approval before acting, which serves as a manual oversight boundary.
  • Capability inventory: The agent has the capability to write files via brew bundle dump, execute package upgrades, and run system-level updates using sudo.
  • Sanitization: There is no automated sanitization of tool output; safety relies on the structural safety rails and human-in-the-loop confirmation steps defined in the instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 21, 2026, 06:29 PM
Security Audit — agent-trust-hub — mac-maintenance