evil-winrm

Pass

Audited by Gen Agent Trust Hub on Jun 23, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONNO_CODE
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by instructing the agent to interpolate untrusted data (target hosts and credentials) directly into shell command templates.
  • Ingestion points: The skill accepts 'Target host' and 'Authorized credential' as inputs in SKILL.md.
  • Boundary markers: No delimiters or instructions to ignore embedded commands are present in the command templates.
  • Capability inventory: The skill documents the use of the evil-winrm tool for remote shell access and file operations (uploading executables, loading scripts).
  • Sanitization: No validation or escaping logic is described for the provided inputs.
  • [COMMAND_EXECUTION]: The documentation provides multiple command-line examples for the evil-winrm utility. This tool is specifically designed for establishing remote shell sessions and executing commands on Windows hosts via the WinRM protocol.
  • [NO_CODE]: The skill consists solely of a markdown documentation file (SKILL.md). It does not include any accompanying scripts (.sh, .py, .js), binaries, or configuration files that would be executed automatically by the agent environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 23, 2026, 02:37 AM
Security Audit — agent-trust-hub — evil-winrm