mdm-device-management
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPRIVILEGE_ESCALATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: Fetches the Homebrew installation script from its official repository (raw.githubusercontent.com/Homebrew/install/HEAD/install.sh). This is a well-known service and a standard method for installing package managers on macOS.- [REMOTE_CODE_EXECUTION]: Executes the Homebrew installation script using bash via command substitution. While this involves executing remote code, the source is a trusted provider and the action is required for the skill's documented purpose of managing software deployments.- [PRIVILEGE_ESCALATION]: Utilizes
sudoto modify system-level security configurations, including enabling FileVault disk encryption (fdesetup) and the application firewall (socketfilterfw). These actions are necessary for the skill's primary function of securing company endpoints.- [INDIRECT_PROMPT_INJECTION]: Identifies an attack surface where the agent processes external device metadata which could potentially contain malicious instructions. - Ingestion points: Device status and query results retrieved through
fleetctlandosquery(SKILL.md). - Boundary markers: Not present in the provided shell script or configuration templates.
- Capability inventory: System configuration changes via
sudo, remote device lock/wipe commands viafleetctl, and package installation viabrewandwinget. - Sanitization: The example onboarding scripts do not demonstrate explicit sanitization or validation of data retrieved from endpoints.
Audit Metadata