Use WebSureQTool from AI agents (MCP server)

WebSureQTool 1.5 lets an AI coding agent use it the way agents use a browser automation library: learn the step vocabulary, write a suite, validate it, run it in a real browser, read structured failures, and fix them.

WHAT THE AGENT CAN DO
- Describe: list every step type with its fields and a working example.
- Write and validate suites and datasets in your workspace, before any browser opens.
- Run suites in Edge, Chrome or Firefox (headless by default) and get per-suite pass/fail, the failing step, the error and a screenshot.
- Run OpenAPI suites, scan a site into draft suites, install starter projects and check saved sign-ins.

CONNECT AN AGENT
The installer ships wsq.exe next to WebSureQTool.exe. Register it as an MCP server named "wsq" with the arguments: mcp --workspace <your workspace folder>. This works in Claude Code, Claude Desktop, VS Code with GitHub Copilot agent mode, Cursor, Windsurf and other MCP clients. The Intelligent Assistant has the exact configuration for each.

WHY IT STAYS TRUSTWORTHY
- The agent writes plain YAML suites you can read and review; nothing is hidden in a model.
- Runs are deterministic and identical to runs a person starts: same engine, reports, saved sign-ins and exit codes.
- The agent only reads and writes inside the workspace you give it, and it can never read saved sign-in files.
- WebSureQTool does not include or call an AI model itself. You bring the agent; WebSureQTool gives it a reliable way to test.

The public guide stays intentionally high level. Open the Intelligent Assistant for click-by-click instructions, learning order, and practice ideas.