macos-dev-setup
Pass
Audited by Gen Agent Trust Hub on Sep 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONPRIVILEGE_ESCALATIONEXTERNAL_DOWNLOADSDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes multiple shell commands to inventory the system and verify the setup.
- Evidence:
sw_vers,uname -m,xcode-select -p,brew --version,df -h /,brew doctor,node -v,python3 --version,docker info. - [PRIVILEGE_ESCALATION]: The skill orchestrates tasks that require elevated permissions.
- Evidence: Configuration of FileVault, firewall, and Touch ID for
sudorequires administrative access. - [EXTERNAL_DOWNLOADS]: The skill manages the installation of various software packages from external sources.
- Evidence: Directs the installation of development tools, CLI utilities, and runtimes, primarily via Homebrew.
- [DATA_EXFILTRATION]: The skill accesses system and user metadata for inventory and connectivity testing.
- Evidence: Reads
git config --global user.email, checks for the presence of SSH keys usingls ~/.ssh/id_*, and performs a connectivity check to a well-known service usingssh -T git@github.com. - [INDIRECT_PROMPT_INJECTION]: The skill has a user-input surface for plan customization.
- Ingestion points: User confirmation and track selection in
SKILL.md(Step 2). - Boundary markers: Includes a mandatory confirmation step ("Propose the plan, ask twice, then go").
- Capability inventory: Extensive command execution capabilities throughout the workflow.
- Sanitization: Not explicitly defined in this orchestrator skill; relies on sub-skill safety rails.
Audit Metadata