I am looking for an AI automation expert to build an autonomous QA agent (using OpenClaw or your preferred AI agent framework) for a Web-based HRIS platform. The agent's goal is to act as a "Chaos Monkey" / Exploratory Tester. It should simulate a busy HR manager in a 100-employee company by continuously and creatively executing various HR scenarios (hiring, terminating, leave management, changing roles, etc.). The sole purpose of these actions is to find edge cases, system bugs, and UX flaws. Platform Context: Environment: This is a Web-based HR platform. You do not need to host or set up the HR system locally. I will provide the Staging/Testing URL and the necessary login credentials for the agent to access the system directly via the internet. How it should work: Total Autonomy: The agent should independently decide what scenarios to run. I do not need to approve its actions in the test environment—it can hire and fire simulated employees as much as it wants. Alerts & Reporting: The agent must ONLY contact me when it discovers a genuine bug, a system crash, or has a logical proposal to improve the platform. Do not spam me with successful test reports. Reporting Format: When an issue or improvement is found, the agent must send an Email formatted as a "TASK". This email must include: A concise text description of the bug or improvement. A screenshot of the web page exactly when and where the issue occurred. Technical Freedom: You are free to choose the infrastructure, browser automation scripts, and AI orchestration methods that you believe are best suited for this task. I only care about the result. Definition of Done (DoD): The project will be considered 100% complete and ready for payment when the following criteria are met: Deployment: The agent is deployed (e.g., on a VPS or cloud service of your choice) and runs continuously in the background without requiring my manual intervention to start/stop. Scenario Execution: The agent successfully and autonomously performs complex, multi-step actions in the web HR portal (e.g., creating an employee, modifying their data, and later terminating them). Bug Reporting (The Core Goal): The agent correctly identifies at least one valid system error or UI/UX anomaly and emails a clear "TASK" report to my designated email address. Visual Evidence: The emailed bug report MUST include an automated screenshot showing the exact state of the UI where the issue occurred. Silence on Success: The agent successfully runs background operations without pinging or notifying me for standard, successful actions.