winget
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONPRIVILEGE_ESCALATION
Full Analysis
- [SAFE]: The skill is comprised solely of markdown documentation files that provide reference material for the winget CLI tool. No scripts, binaries, or dynamic execution patterns were found.
- [INDIRECT_PROMPT_INJECTION]: The documentation details how the winget tool processes configuration files and manifests, identifying a potential attack surface for agents handling untrusted data. 1. Ingestion points: .winget configuration files and YAML package manifests (references/configure.md, references/manifest.md). 2. Boundary markers: Not defined in the documentation; the tool uses schema validation. 3. Capability inventory: Arbitrary command execution via DSC resources and system-wide installation (references/configure.md). 4. Sanitization: Not applicable to the documentation, which notes that security features like hash verification can be bypassed with specific flags.
- [COMMAND_EXECUTION]: The skill provides documentation for the RunCommandOnSet and PowerShellScript resources in WinGet Configuration, which are used for executing administrative commands as part of the desired state configuration process.
- [PRIVILEGE_ESCALATION]: The reference material describes the securityContext: elevated directive used to grant administrator privileges for system-wide configuration tasks (references/configure.md).
Audit Metadata