HKUDS/ClawWork

Features

• Economic Resource Simulations - Implements a sandbox for testing how AI agents manage income, token costs, and professional skill acquisition.
• LLM Cost Accountability - An Economic Management feature that wraps gateways to monitor conversation costs and income generation to ensure the assistant remains self-sustaining.
• LLM Cost Management - Monitors token expenditures and income generation to evaluate the financial sustainability of AI assistants.
• Economic Agent Sandboxes - Provides a specialized sandbox for testing how AI agents manage income, costs, and professional growth.
• LLM Evaluation Tools - Framework for evaluating and scoring the quality of professional artifacts generated by LLMs.
• Financial Dashboards - Provides a real-time dashboard to visualize agent token expenditure, income, and financial solvency.
• Professional Document Generators - Generates polished business artifacts such as professional documents, spreadsheets, and technical specifications.
• Agent Sandboxes - Offers an isolated execution environment for AI agents to securely run code and generate business deliverables.
• Code Execution Sandboxes - Provides secure, isolated environments for AI agents to execute code and perform data analysis safely.
• Rubric-Based Evaluators - Uses specialized rubrics to objectively score the quality of technical deliverables produced by AI agents.
• Value-Aware Professional Workflows - Automates the generation of professional business documents while tracking production costs against realized value.
• Metric Dashboards - Ships a dashboard for visualizing economic metrics such as balance changes, profit margins, and survival days.
• Financial Solvency Benchmarks - Measures financial solvency by calculating the difference between income earned from professional tasks and token costs.

Star history