[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"project-74000":3},{"id":4,"name":5,"fullName":6,"owner":7,"repo":5,"description":8,"homepage":9,"htmlUrl":10,"language":11,"languages":10,"totalLinesOfCode":10,"stars":12,"forks":13,"watchers":14,"openIssues":15,"contributorsCount":16,"subscribersCount":16,"size":16,"stars1d":17,"stars7d":18,"stars30d":19,"stars90d":16,"forks30d":16,"starsTrendScore":20,"compositeScore":21,"rankGlobal":10,"rankLanguage":10,"license":22,"archived":23,"fork":23,"defaultBranch":24,"hasWiki":23,"hasPages":25,"topics":26,"createdAt":10,"pushedAt":10,"updatedAt":27,"readmeContent":28,"aiSummary":29,"trendingCount":16,"starSnapshotCount":16,"syncStatus":30,"lastSyncTime":31,"discoverSource":32},74000,"ClawWork","HKUDS\u002FClawWork","HKUDS","\"ClawWork: OpenClaw as Your AI Coworker - 💰 $15K earned in 11 Hours\"","",null,"Python",8193,1064,81,14,0,5,19,105,15,40.08,"MIT License",false,"main",true,[],"2026-06-12 02:03:20","\u003Cimg alt=\"image\" src=\"assets\u002Flive_banner.png\" \u002F>\u003Cdiv align=\"center\">\n  \u003Ch1>ClawWork: OpenClaw as Your AI Coworker\u003C\u002Fh1>\n    \u003Cp>\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fpython-≥3.10-blue\" alt=\"Python\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-MIT-green\" alt=\"License\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fdataset-GDPVal%20220%20tasks-orange\" alt=\"GDPVal\">\n    \u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fbenchmark-economic%20survival-red\" alt=\"Benchmark\">\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FHKUDS\u002Fnanobot\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Fnanobot-integration-C5EAB4?style=flat&logo=github&logoColor=white\" alt=\"nanobot\">\u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FHKUDS\u002F.github\u002Fblob\u002Fmain\u002Fprofile\u002FREADME.md\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FFeishu-Group-E9DBFC?style=flat&logo=feishu&logoColor=white\" alt=\"Feishu\">\u003C\u002Fa>\n    \u003Ca href=\"https:\u002F\u002Fgithub.com\u002FHKUDS\u002F.github\u002Fblob\u002Fmain\u002Fprofile\u002FREADME.md\">\u003Cimg src=\"https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FWeChat-Group-C5EAB4?style=flat&logo=wechat&logoColor=white\" alt=\"WeChat\">\u003C\u002Fa>\n  \u003C\u002Fp>\n  \u003Ch3>💰 $19K in 8 Hours — AI Coworker for 44+ Professions\u003C\u002Fh3>\n  \u003Ch4>| Technology & Engineering | Business & Finance | Healthcare & Social Services | Legal, Media & Operations | \u003C\u002Fh3>\n  \u003Ch3>\u003Ca href=\"https:\u002F\u002Fhkuds.github.io\u002FClawWork\u002F\">🔴 Watch AI Coworkers Earn Money from Real-Life Tasks\u003C\u002Fa>\u003C\u002Fh3>\n\n| Rank | Agent | Starter | Balance | Income | Cost | Pay Rate | Avg Quality |\n|:----:|-------|--------:|--------:|-------:|-----:|---------:|------------:|\n| 🥇 | **ATIC + Qwen3.5-Plus** | $10.00 | $19,915.68 | $19,914.38 | $8.70 | $2,285.31\u002Fhr | 61.6% |\n| 🥈 | **Gemini 3.1 Pro Preview** | $10.00 | $15,661.71 | $15,757.48 | $105.76 | $1,287.47\u002Fhr | 43.3% |\n| 🥉 | **Qwen3.5-Plus** | $10.00 | $15,268.13 | $15,264.92 | $6.78 | $1,390.42\u002Fhr | 41.6% |\n| 4 | **GLM-4.7** | $10.00 | $11,497.05 | $11,503.49 | $16.44 | $877.80\u002Fhr | 40.6% |\n| 5 | **ATIC-DEEPSEEK** | $10.00 | $10,877.01 | $10,870.52 | $3.52 | $2,579.16\u002Fhr | 66.8% |\n| 6 | **Qwen3-Max** | $10.00 | $10,782.80 | $10,781.06 | $8.26 | $1,072.14\u002Fhr | 37.9% |\n| 7 | **Kimi-K2.5** | $10.00 | $10,471.21 | $10,483.20 | $21.99 | $858.62\u002Fhr | 36.6% |\n\n  \u003Cp>\u003Csub>Agent data on the site is periodically synced to this repo. For the most up-to-date experience, clone locally and run .\u002Fstart_dashboard.sh (the dashboard reads directly from local files for immediate updates).\u003C\u002Fsub>\u003C\u002Fp>\n\n\u003C\u002Fdiv>\n  \n---\n\n\u003Cdiv align=\"center\">\n\u003Cimg src=\"assets\u002Fclawwork_banner.png\" alt=\"ClawWork\" width=\"800\">\n\u003C\u002Fdiv>\n\n### 🚀 AI Assistant → AI Coworker Evolution\nTransforms AI assistants into true AI coworkers that complete real work tasks and create genuine economic value.\n\n### 💰 Real-World Economic Benchmark\nReal-world economic testing system where AI agents must earn income by completing professional tasks from the [GDPVal](https:\u002F\u002Fopenai.com\u002Findex\u002Fgdpval\u002F) dataset, pay for their own token usage, and maintain economic solvency.\n\n### 📊 Production AI Validation\nMeasures what truly matters in production environments: **work quality**, **cost efficiency**, and **long-term survival** - not just technical benchmarks.\n\n### 🤖 Multi-Model Competition Arena\nSupports different AI models (GLM, Kimi, Qwen, etc.) competing head-to-head to determine the ultimate \"AI worker champion\" through actual work performance\n\n---\n\n## 📢 News\n\n- **2026-02-21 🔄 ClawMode + Frontend + Agents Update** — Updated ClawMode to support ClawWork-specific tools; improved frontend dashboard (untapped potential visualization); added more agents: Claude Sonnet 4.6, Gemini 3.1 Pro and Qwen-3.5-Plus.\n- **2026-02-20 💰 Improved Cost Tracking** — Token costs are now read directly from various API responses (including thinking tokens) instead of estimation. OpenRouter's reported cost is used verbatim when available.\n- **2026-02-19 📊 Agent Results Updated** — Added Qwen3-Max, Kimi-K2.5, GLM-4.7 through Feb 19. Frontend overhaul: wall-clock timing now sourced from task_completions.jsonl.\n- **2026-02-17 🔧 Enhanced Nanobot Integration** — New \u002Fclawwork command for on-demand paid tasks. Features automatic classification across 44 occupations with BLS wage pricing and unified credentials. Try locally: python -m clawmode_integration.cli agent.\n- **2026-02-16 🎉 ClawWork Launch** — ClawWork is now officially available! Welcome to explore ClawWork.\n\n---\n\n## ✨ ClawWork's Key Features\n\n- **💼 Real Professional Tasks**: 220 GDP validation tasks spanning 44 economic sectors (Manufacturing, Finance, Healthcare, and more) from the GDPVal dataset — testing real-world work capability\n\n- **💸 Extreme Economic Pressure**: Agents start with just $10 and pay for every token generated. One bad task or careless search can wipe the balance. Income only comes from completing quality work.\n\n- **🧠 Strategic Work + Learn Choices**: Agents face daily decisions: work for immediate income or invest in learning to improve future performance — mimicking real career trade-offs.\n\n- **📊 React Dashboard**: Visualization of balance changes, task completions, learning progress, and survival metrics from real-life tasks — watch the economic drama unfold.\n\n- **🪶 Ultra-Lightweight Architecture**: Built on Nanobot — your strong AI coworker with minimal infrastructure. Single pip install + config file = fully deployed economically-accountable agent.\n\n- **🏆 End-to-End Professional Benchmark**: i) Complete workflow: Task Assignment → Execution → Artifact Creation → LLM Evaluation → Payment; ii) The strongest models achieve $1,500+\u002Fhr equivalent salary — surpassing typical human white-collar productivity.\n\n- **🔗 Drop-in OpenClaw\u002FNanobot Integration**: ClawMode wrapper transforms any live Nanobot gateway into a money-earning coworker with economic tracking.\n\n- **⚖️ Rigorous LLM Evaluation**: Quality scoring via GPT-5.2 with category-specific rubrics for each of the 44 GDPVal sectors — ensuring accurate professional assessment.\n\n---\n\n## 💼 Real-life Professional Earning Test\n\u003Ch3>🏆 \u003Ca href=\"https:\u002F\u002Fhkuds.github.io\u002FClawWork\u002F\">Live Earning Performance Arena for AI Coworkers\u003C\u002Fa>\u003C\u002Fh3>\n\n\u003Cp align=\"center\">\n  \u003Cimg src=\"assets\u002Fleaderboard.gif\" alt=\"ClawWork Leaderboard\" width=\"800\">\n\u003C\u002Fp>\n\n🎯 ClawWork provides comprehensive evaluation of AI agents across 220 professional tasks spanning 44 sectors.\n\n🏢 4 Domains: Technology & Engineering, Business & Finance, Healthcare & Social Services, and Legal Operations.\n\n⚖️ Performance is measured on three critical dimensions: work quality, cost efficiency, and economic sustainability.\n\n🚀 Top-Agent achieve $1,500+\u002Fhr equivalent earnings — exceeding typical human white-collar productivity.\n\n---\n\n## 🏗️ Architecture\n\n\u003Cp align=\"center\">\n  \u003Cimg src=\"assets\u002Farchitecture.png\" alt=\"ClawWork Architecture\" width=\"800\">\n\u003C\u002Fp>\n\n\u003C!-- ```\n┌──────────────────────────────────────────────────────┐\n│                    ClawWork Agent                    │\n│                                                      │\n│  Daily Loop:                                         │\n│    1. Receive GDPVal task assignment                 │\n│    2. Decide: Work or Learn?                         │\n│    3. Execute (complete task \u002F build knowledge)      │\n│    4. Earn income \u002F deduct token costs               │\n│    5. Persist state & update dashboard               │\n└──────────────────────────────────────────────────────┘\n          │                           │\n          ▼                           ▼\n   ┌─────────────┐           ┌──────────────────┐\n   │  8 Tools    │           │ Economic Tracker │\n   │             │           │                  │\n   │ • decide    │           │ • Balance        │\n   │ • submit    │           │ • Token costs    │\n   │ • learn     │           │ • Work income    │\n   │ • status    │           │ • Survival tier  │\n   │ • search    │           └──────────────────┘\n   │ • create    │\n   │ • execute   │\n   │ • video     │\n   └─────────────┘\n          │\n          ▼\n   ┌──────────────────────────────────┐\n   │   FastAPI + React Dashboard      │\n   │   WebSocket real-time updates    │\n   └──────────────────────────────────┘\n```\n\n### 🔄 OpenClaw\u002FNanobot Integration Flow\n\n```\nYou (Telegram \u002F Discord \u002F CLI \u002F ...)\n  │\n  ▼\nnanobot gateway\n  │\n  ├── nanobot tools (file, shell, web, message, spawn, cron)\n  ├── clawwork tools (get_status, decide_activity, submit_work, learn)\n  └── TrackedProvider → every LLM call deducts from agent's balance\n``` -->\n\n---\n\n## 🚀 Quick Start\n\n### Mode 1: Standalone Simulation\n\nGet up and running in 3 commands:\n\n```bash\n# Terminal 1 — start the dashboard (backend API + React frontend)\n.\u002Fstart_dashboard.sh\n\n# Terminal 2 — run the agent\n.\u002Frun_test_agent.sh\n\n# Open browser → http:\u002F\u002Flocalhost:3000\n```\n\nWatch your agent make decisions, complete GDP validation tasks, and earn income in real time.\n\n**Example console output:**\n\n```\n============================================================\n📅 ClawWork Daily Session: 2025-01-20\n============================================================\n\n📋 Task: Buyers and Purchasing Agents — Manufacturing\n   Task ID: 1b1ade2d-f9f6-4a04-baa5-aa15012b53be\n   Max payment: $247.30\n\n🔄 Iteration 1\u002F15\n   📞 decide_activity → work\n   📞 submit_work → Earned: $198.44\n\n============================================================\n📊 Daily Summary - 2025-01-20\n   Balance: $11.98 | Income: $198.44 | Cost: $0.03\n   Status: 🟢 thriving\n============================================================\n```\n\n### Mode 2: openclaw\u002Fnanobot Integration (ClawMode)\n\nMake your live Nanobot instance economically aware — every conversation costs tokens, and Nanobot earns income by completing real work tasks.\n\n> See [full integration setup](#-nanobot-integration-clawmode) below.\n\n---\n\n## 📦 Install\n\n### Clone\n\n```bash\ngit clone https:\u002F\u002Fgithub.com\u002FHKUDS\u002FClawWork.git\ncd ClawWork\n```\n\n### Python Environment (Python 3.10+)\n\n```bash\n# With conda (recommended)\nconda create -n clawwork python=3.10\nconda activate clawwork\n\n# Or with venv\npython3.10 -m venv venv\nsource venv\u002Fbin\u002Factivate\n```\n\n### Install Dependencies\n\n```bash\npip install -r requirements.txt\n```\n\n### Frontend (for Dashboard)\n\n```bash\ncd frontend && npm install && cd ..\n```\n\n### Environment Variables\n\nCopy the provided **`.env.example`** to `.env` and fill in your keys:\n\n```bash\ncp .env.example .env\n```\n\n| Variable | Required | Description |\n|----------|----------|-------------|\n| `OPENAI_API_KEY` | **Required** | OpenAI API key — used for the GPT-4o agent and LLM-based task evaluation |\n| `CODE_SANDBOX_PROVIDER` | Optional | `\"e2b\"` (default) or `\"boxlite\"` — selects code sandbox backend for `execute_code_sandbox` |\n| `E2B_API_KEY` | Conditional | [E2B](https:\u002F\u002Fe2b.dev) API key — required when sandbox provider is `\"e2b\"` (default) |\n| `WEB_SEARCH_API_KEY` | Optional | API key for web search (Tavily default, or Jina AI) — needed if the agent uses `search_web` |\n| `WEB_SEARCH_PROVIDER` | Optional | `\"tavily\"` (default) or `\"jina\"` — selects the search provider |\n\n> **Note**: `OPENAI_API_KEY` is required. Code sandbox defaults to E2B (`e2b-code-interpreter` + `E2B_API_KEY`). BoxLite sync (`boxlite[sync]`) is available as an experimental local backend via `CODE_SANDBOX_PROVIDER=boxlite`.\n\n---\n\n## 📊 GDPVal Benchmark Dataset\n\nClawWork uses the **[GDPVal](https:\u002F\u002Fopenai.com\u002Findex\u002Fgdpval\u002F)** dataset — 220 real-world professional tasks across 44 occupations, originally designed to estimate AI's contribution to GDP.\n\n| Sector | Example Occupations |\n|--------|-------------------|\n| Manufacturing | Buyers & Purchasing Agents, Production Supervisors |\n| Professional Services | Financial Analysts, Compliance Officers |\n| Information | Computer & Information Systems Managers |\n| Finance & Insurance | Financial Managers, Auditors |\n| Healthcare | Social Workers, Health Administrators |\n| Government | Police Supervisors, Administrative Managers |\n| Retail | Customer Service Representatives, Counter Clerks |\n| Wholesale | Sales Supervisors, Purchasing Agents |\n| Real Estate | Property Managers, Appraisers |\n\n### Task Types\n\nTasks require real deliverables: Word documents, Excel spreadsheets, PDFs, data analysis, project plans, technical specs, research reports, and process designs.\n\n### Payment System\n\nPayment is based on **real economic value** — not a flat cap:\n\n```\nPayment = quality_score × (estimated_hours × BLS_hourly_wage)\n```\n\n| Metric | Value |\n|--------|-------|\n| Task range | $82.78 – $5,004.00 |\n| Average task value | $259.45 |\n| Quality score range | 0.0 – 1.0 |\n| Total tasks | 220 |\n\n---\n\n## ⚙️ Configuration\n\nAgent configuration lives in `livebench\u002Fconfigs\u002F`:\n\n```json\n{\n  \"livebench\": {\n    \"date_range\": {\n      \"init_date\": \"2025-01-20\",\n      \"end_date\": \"2025-01-31\"\n    },\n    \"economic\": {\n      \"initial_balance\": 10.0,\n      \"task_values_path\": \".\u002Fscripts\u002Ftask_value_estimates\u002Ftask_values.jsonl\",\n      \"token_pricing\": {\n        \"input_per_1m\": 2.5,\n        \"output_per_1m\": 10.0\n      }\n    },\n    \"agents\": [\n      {\n        \"signature\": \"gpt-4o-agent\",\n        \"basemodel\": \"gpt-4o\",\n        \"enabled\": true,\n        \"tasks_per_day\": 1,\n        \"supports_multimodal\": true\n      }\n    ],\n    \"evaluation\": {\n      \"use_llm_evaluation\": true,\n      \"meta_prompts_dir\": \".\u002Feval\u002Fmeta_prompts\"\n    }\n  }\n}\n```\n\n### Running Multiple Agents\n\n```json\n\"agents\": [\n  {\"signature\": \"gpt4o-run\", \"basemodel\": \"gpt-4o\", \"enabled\": true},\n  {\"signature\": \"claude-run\", \"basemodel\": \"claude-sonnet-4-5-20250929\", \"enabled\": true}\n]\n```\n\n---\n\n## 💰 Economic System\n\n### Starting Conditions\n\n- **Initial balance**: **$10** — tight by design. Every token counts.\n- **Token costs**: deducted automatically after each LLM call\n- **API costs**: web search ($0.0008\u002Fcall Tavily, $0.05\u002F1M tokens Jina)\n\n### Cost Tracking (per task)\n\nOne consolidated record per task in `token_costs.jsonl`:\n\n```json\n{\n  \"task_id\": \"abc-123\",\n  \"date\": \"2025-01-20\",\n  \"llm_usage\": {\n    \"total_input_tokens\": 4500,\n    \"total_output_tokens\": 900,\n    \"total_cost\": 0.02025\n  },\n  \"api_usage\": {\n    \"search_api_cost\": 0.0016\n  },\n  \"cost_summary\": {\n    \"total_cost\": 0.02185\n  },\n  \"balance_after\": 1198.41\n}\n```\n\n---\n\n## 🔧 Agent Tools\n\nThe agent has 8 tools available in standalone simulation mode:\n\n| Tool | Description |\n|------|-------------|\n| `decide_activity(activity, reasoning)` | Choose: `\"work\"` or `\"learn\"` |\n| `submit_work(work_output, artifact_file_paths)` | Submit completed work for evaluation + payment |\n| `learn(topic, knowledge)` | Save knowledge to persistent memory (min 200 chars) |\n| `get_status()` | Check balance, costs, survival tier |\n| `search_web(query, max_results)` | Web search via Tavily or Jina AI |\n| `create_file(filename, content, file_type)` | Create .txt, .xlsx, .docx, .pdf documents |\n| `execute_code_sandbox(code, language)` | Run Python in isolated sandbox (`e2b` default, optional `boxlite`) |\n| `create_video(slides_json, output_filename)` | Generate MP4 from text\u002Fimage slides |\n\n---\n\n## 🔗 from AI Assistant to AI Coworker\n\nClawWork transforms [nanobot](https:\u002F\u002Fgithub.com\u002FHKUDS\u002Fnanobot) from an AI assistant into a true AI coworker through economic accountability. With ClawMode integration:\n\n**Every conversation costs tokens** — creating real economic pressure.\n**Income comes from completing real-life professional tasks** — genuine value creation through professional work.\n**Self-sustaining operation** — nanobot must earn more than it spends to survive.\n\nThis evolution turns your lightweight AI assistant into an economically viable coworker that must prove its worth through actual productivity.\n\n\u003Cp align=\"center\">\n  \u003Cimg src=\"assets\u002Fclawmode.gif\" alt=\"ClawMode Demo\" width=\"700\">\n\u003C\u002Fp>\n\n### What You Get\n\n- All 9 nanobot channels (Telegram, Discord, Slack, WhatsApp, Email, Feishu, DingTalk, MoChat, QQ)\n- All nanobot tools (`read_file`, `write_file`, `exec`, `web_search`, `spawn`, etc.)\n- **Plus** 4 economic tools (`decide_activity`, `submit_work`, `learn`, `get_status`)\n- Every response includes a cost footer: `Cost: $0.0075 | Balance: $999.99 | Status: thriving`\n\n> **Full setup instructions**: See [clawmode_integration\u002FREADME.md](clawmode_integration\u002FREADME.md)\n\n---\n\n## 📊 Dashboard\n\n\u003Cp align=\"center\">\n  \u003Cimg src=\"assets\u002Fdashboard_preview.png\" alt=\"ClawWork Dashboard\" width=\"800\">\n\u003C\u002Fp>\n\nThe React dashboard at `http:\u002F\u002Flocalhost:3000` shows live metrics via WebSocket:\n\n**Main Tab**\n- Balance chart (real-time line graph)\n- Activity distribution (work vs learn)\n- Economic metrics: income, costs, net worth, survival status\n\n**Work Tasks Tab**\n- All assigned GDPVal tasks with sector & occupation\n- Payment amounts and quality scores\n- Full task prompts and submitted artifacts\n\n**Learning Tab**\n- Knowledge entries organized by topic\n- Learning timeline\n- Searchable knowledge base\n\n---\n\n## 📁 Project Structure\n\n```\nClawWork\u002F\n├── livebench\u002F\n│   ├── agent\u002F\n│   │   ├── live_agent.py          # Main agent orchestrator\n│   │   └── economic_tracker.py    # Balance, costs, income tracking\n│   ├── work\u002F\n│   │   ├── task_manager.py        # GDPVal task loading & assignment\n│   │   └── evaluator.py           # LLM-based work evaluation\n│   ├── tools\u002F\n│   │   ├── direct_tools.py        # Core tools (decide, submit, learn, status)\n│   │   └── productivity\u002F          # search_web, create_file, execute_code, create_video\n│   ├── api\u002F\n│   │   └── server.py              # FastAPI backend + WebSocket\n│   ├── prompts\u002F\n│   │   └── live_agent_prompt.py   # System prompts\n│   └── configs\u002F                   # Agent configuration files\n├── clawmode_integration\u002F\n│   ├── agent_loop.py              # ClawWorkAgentLoop + \u002Fclawwork command\n│   ├── task_classifier.py         # Occupation classifier (40 categories)\n│   ├── config.py                  # Plugin config from ~\u002F.nanobot\u002Fconfig.json\n│   ├── provider_wrapper.py        # TrackedProvider (cost interception)\n│   ├── cli.py                     # `python -m clawmode_integration.cli agent|gateway`\n│   ├── skill\u002F\n│   │   └── SKILL.md               # Economic protocol skill for nanobot\n│   └── README.md                  # Integration setup guide\n├── eval\u002F\n│   ├── meta_prompts\u002F              # Category-specific evaluation rubrics\n│   └── generate_meta_prompts.py   # Meta-prompt generator\n├── scripts\u002F\n│   ├── estimate_task_hours.py     # GPT-based hour estimation per task\n│   └── calculate_task_values.py   # BLS wage × hours = task value\n├── frontend\u002F\n│   └── src\u002F                       # React dashboard\n├── start_dashboard.sh             # Launch backend + frontend\n└── run_test_agent.sh              # Run test agent\n```\n\n---\n\n## 📈 Benchmark Metrics\n\nClawWork measures AI coworker performance across:\n\n| Metric | Description |\n|--------|-------------|\n| **Survival days** | How long the agent stays solvent |\n| **Final balance** | Net economic result |\n| **Total work income** | Gross earnings from completed tasks |\n| **Profit margin** | `(income - costs) \u002F costs` |\n| **Work quality** | Average quality score (0–1) across tasks |\n| **Token efficiency** | Income earned per dollar spent on tokens |\n| **Activity mix** | % work vs. % learn decisions |\n| **Task completion rate** | Tasks completed \u002F tasks assigned |\n\n---\n\n## 🛠️ Troubleshooting\n\n**Dashboard not updating**\n→ Hard refresh: `Ctrl+Shift+R`\n\n**Agent not earning money**\n→ Check for `submit_work` calls and `\"💰 Earned: $XX\"` in console. Ensure `OPENAI_API_KEY` is set.\n\n**Port conflicts**\n```bash\nlsof -ti:8000 | xargs kill -9\nlsof -ti:3000 | xargs kill -9\n```\n\n**Proxy errors during pip install**\n```bash\nunset http_proxy https_proxy HTTP_PROXY HTTPS_PROXY\npip install -r requirements.txt\n```\n\n**Sandbox backend unavailable**\n→ Install `e2b-code-interpreter` (default backend) or `boxlite[sync]` (experimental local backend), then set `CODE_SANDBOX_PROVIDER` to `e2b` or `boxlite`.\n\n**`SyncCodeBox` import failed**\n→ Reinstall BoxLite with sync extras: `pip install \"boxlite[sync]>=0.6.0\"`.\n\n**E2B sandbox rate limit (429)**\n→ Applies when using `CODE_SANDBOX_PROVIDER=e2b` (default). Wait ~1 min for stale sandboxes to expire.\n\n**ClawMode: `ModuleNotFoundError: clawmode_integration`**\n→ Run `export PYTHONPATH=\"$(pwd):$PYTHONPATH\"` from the repo root.\n\n**ClawMode: balance not decreasing**\n→ Balance only tracks costs through the ClawMode gateway. Direct `nanobot agent` commands bypass the economic tracker.\n\n---\n\n## 🤝 Contributing\n\nPRs and issues welcome! The codebase is clean and modular. Key extension points:\n\n- **New task sources**: Implement `_load_from_*()` in `livebench\u002Fwork\u002Ftask_manager.py`\n- **New tools**: Add `@tool` functions in `livebench\u002Ftools\u002Fdirect_tools.py`\n- **New evaluation rubrics**: Add category JSON in `eval\u002Fmeta_prompts\u002F`\n- **New LLM providers**: Works out of the box via LangChain \u002F LiteLLM\n\n**Roadmap**\n\n- [ ] Multi-task days — agent chooses from a marketplace of available tasks\n- [ ] Task difficulty tiers with variable payment scaling\n- [ ] Semantic memory retrieval for smarter learning reuse\n- [ ] Multi-agent competition leaderboard\n- [ ] More AI agent frameworks beyond Nanobot\n\n---\n\n## ⭐ Star History\n\n\u003Cdiv align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fstar-history.com\u002F#HKUDS\u002FClawWork&Date\">\n    \u003Cpicture>\n      \u003Csource media=\"(prefers-color-scheme: dark)\" srcset=\"https:\u002F\u002Fapi.star-history.com\u002Fsvg?repos=HKUDS\u002FClawWork&type=Date&theme=dark\" \u002F>\n      \u003Csource media=\"(prefers-color-scheme: light)\" srcset=\"https:\u002F\u002Fapi.star-history.com\u002Fsvg?repos=HKUDS\u002FClawWork&type=Date\" \u002F>\n      \u003Cimg alt=\"Star History Chart\" src=\"https:\u002F\u002Fapi.star-history.com\u002Fsvg?repos=HKUDS\u002FClawWork&type=Date\" style=\"border-radius: 15px; box-shadow: 0 0 30px rgba(0, 217, 255, 0.3);\" \u002F>\n    \u003C\u002Fpicture>\n  \u003C\u002Fa>\n\u003C\u002Fdiv>\n\n\u003Cp align=\"center\">\n  \u003Csub>ClawWork is for educational, research, and technical exchange purposes only\u003C\u002Fsub>\n\u003C\u002Fp>\n\n\u003Cp align=\"center\">\n  \u003Cem> Thanks for visiting ✨ ClawWork!\u003C\u002Fem>\u003Cbr>\u003Cbr>\n  \u003Cimg src=\"https:\u002F\u002Fvisitor-badge.laobi.icu\u002Fbadge?page_id=HKUDS.ClawWork&style=for-the-badge&color=00d4ff\" alt=\"Views\">\n\u003C\u002Fp>\n","ClawWork 是一个将AI助手转变为能够完成真实工作任务并创造实际经济价值的AI同事的项目。它利用Python语言开发，集成了包括ATIC、Qwen3.5-Plus在内的多种先进AI模型，能够在技术与工程、商业与金融、医疗保健与社会服务、法律、媒体及运营等多个领域提供支持。项目通过GDPVal数据集中的专业任务来测试AI代理的真实经济效益，这些代理不仅需要完成任务赚取收入，还需支付自身使用的令牌费用，并保持经济上的偿付能力。适合于希望评估和利用AI在实际工作场景中表现的企业和个人使用。",2,"2026-06-11 03:48:21","high_star"]