windows-qa-engineer

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The prompt's workflows (e.g., the login example) explicitly show putting plaintext credentials into tool calls like set_edit_text, which requires the agent to emit secret values verbatim in its outputs.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The skill includes a runtime auto-setup/installer (scripts/skill_installer.py and install.yaml) that will git clone and pip-install https://github.com/microsoft/UFO.git (and related packages) when MCP tools are missing, which fetches and executes remote code that the skill depends on, so this external URL is a high-confidence runtime code-execution dependency.