Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →This skill provides tooling and documentation to run, inspect, and extend the multi-agent LangGraph trading workflow in OpenClaw environments.
.claude/skills/olaxbt-ai-market-maker-skill/SKILL.md| Test case | Without → With | Effect | Δ tokens | Δ turns |
|---|---|---|---|---|
| case-01 | ✗→✓ | ▲ Improved | 53% | 0% |
| case-03 | ✗→✓ | ▲ Improved | 24% | 0% |
| case-05 | ✗→✓ | ▲ Improved | 14% | 0% |
| case-06 | ✗→✓ | ▲ Improved | 7% | 0% |
| case-08 | ✗→✓ | ▲ Improved | -1% | 0% |
This skill provides tooling and documentation to run, inspect, and extend the multi-agent LangGraph trading workflow in OpenClaw environments.
It includes a hard Risk Guard veto before any execution, structured tracing for transparency, and dedicated OpenClaw integration tools.
bash# From OpenClaw claw install https://github.com/olaxbt/ai-market-maker # Or locally git clone https://github.com/olaxbt/ai-market-maker.git cd ai-market-maker claw skill install ./openclaw
bash# Check dependencies ./openclaw/scripts/verify_installation.sh # Or python3 openclaw/scripts/claw_runner.py --verify
bash# Run backtest with default settings python3 openclaw/scripts/claw_runner.py --backtest # Paper trading python3 openclaw/scripts/claw_runner.py --paper --ticker BTC/USDT # Custom backtest python3 openclaw/scripts/claw_runner.py --backtest --symbols "BTC/USDT,ETH/USDT" --steps 150
The default settings use multiple symbols and conservative risk parameters:
Trade count: 17
Total return: 14.95%
Excess return vs BTC buy & hold: +30.25%
Sharpe ratio: 1.79
Maximum drawdown: 11.84%
Win rate: 62.5%These results are based on 100 days of historical data across BTC, ETH, and SOL, with full benchmark comparison and risk event logging.
bash# Paper trading with custom ticker claw run ai-market-maker --paper --ticker ETH/USDT # Backtesting with multiple symbols claw run ai-market-maker --backtest --symbols "BTC/USDT,ETH/USDT" --steps 150 # Installation verification claw run ai-market-maker --verify
Problem: ModuleNotFoundError: No module named 'talib' Solution:
bash# Recommended for environments without sudo conda install -y ta-lib -c conda-forge # Alternative: source compilation wget http://prdownloads.sourceforge.net/ta-lib/ta-lib-0.4.0-src.tar.gz tar -xzf ta-lib-0.4.0-src.tar.gz cd ta-lib/ ./configure --prefix=$HOME/.local make make install export LD_LIBRARY_PATH=$HOME/.local/lib:$LD_LIBRARY_PATH pip install ta-lib
Problem: ModuleNotFoundError: No module named 'agents.market_scanner' Solution:
bash# Install in development mode pip install -e . # Or set Python path export PYTHONPATH=/path/to/ai-market-maker/src:$PYTHONPATH
Problem: 429 Too Many Requests Solution:
.env for production useProblem: Environment variables not set Solution:
bash# Automatic configuration via claw_runner.py # Manual override: export NEXUS_API_KEY=your_key export AIMM_DESK_STRATEGY_PRESET=default
The repo exposes a lightweight, mostly read-only HTTP API:
GET /runs/latest → Latest run dataGET /runs/{run_id}/payload → Full payload of a runGET /runs/{run_id}/events → Events and tracesGET /pm/portfolio-health → Portfolio summaryGET /backtests → List backtest runsAIMM_API_KEY and AIMM_AUTH_SECRET are always required (no shipped default)..secrets/.GET /health require x-api-key (or Authorization: Bearer with the same key).AIMM_CORS_ORIGINS to your dashboard origin.| Area | Location | Purpose | |------|----------|---------| | OpenClaw Runner | openclaw/scripts/claw_runner.py | Main entry point | | Installation Verifier | openclaw/scripts/verify_installation.sh | Dependency checker | | Skill Manifest | openclaw/manifest.json | OpenClaw skill definition | | Usage Examples | openclaw/examples/claw_usage.md | Usage guides | | Main Workflow | src/main.py | Core trading logic | | Agent System | src/agents/ | 7 trading desks | | Web Dashboard | web/ | Next.js monitoring UI |
STRATEGY_INTERVAL_SEC for lower resource usageWe welcome contributions! Please read the main CONTRIBUTING.md first.
bash# 1. Fork the repository # 2. Create a feature branch git checkout -b feature/improvement # 3. Make your changes # 4. Test with verification script ./openclaw/scripts/verify_installation.sh # 5. Submit Pull Request
README.mdopenclaw/examples/claw_usage.mdopenclaw/examples/korean_guide.mddocs/ directorydocs/ directory and openclaw/examples/Version: 1.0.0 Last Updated: 2026-04-17
| Case | Status | Duration (ms) | Turns | Tokens | Tool calls | ||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Without | With | Δ | Without | With | Δ | Without | With | Δ | Without | With | Δ | ||
case-01 | fail→pass | 39,834 | 33,279 | -16% | 1 | 1 | 0% | 1,426 | 2,178 | +53% | 0 | 0 | — |
case-02 | pass→pass | 37,319 | 18,157 | -51% | 1 | 1 | 0% | 1,370 | 2,181 | +59% | 0 | 0 | — |
case-03 | fail→pass | 38,556 | 3,146 | -92% | 1 | 1 | 0% | 1,702 | 2,103 | +24% | 0 | 0 | — |
case-04 | pass→pass | 6,333 | 3,294 | -48% | 1 | 1 | 0% | 1,132 | 2,177 | +92% | 0 | 0 | — |
case-05 | fail→pass | 20,235 | 3,533 | -83% | 1 | 1 | 0% | 1,965 | 2,234 | +14% | 0 | 0 | — |
case-06 | fail→pass | 11,594 | 2,346 | -80% | 1 | 1 | 0% | 1,827 | 1,948 | +7% | 0 | 0 | — |
case-07 | pass→pass | 17,324 | 2,409 | -86% | 1 | 1 | 0% | 2,794 | 1,962 | -30% | 0 | 0 | — |
case-08 | fail→pass | 24,928 | 2,219 | -91% | 1 | 1 | 0% | 1,926 | 1,913 | -1% | 0 | 0 | — |
case-09 | fail→pass | 47,700 | 2,293 | -95% | 1 | 1 | 0% | 3,012 | 1,928 | -36% | 0 | 0 | — |
case-10 | fail→pass | 17,373 | 2,108 | -88% | 1 | 1 | 0% | 3,128 | 1,880 | -40% | 0 | 0 | — |
case-11 | fail→pass | 11,249 | 2,209 | -80% | 1 | 1 | 0% | 1,881 | 1,972 | +5% | 0 | 0 | — |
case-12 | pass→pass | 9,775 | 1,990 | -80% | 1 | 1 | 0% | 1,579 | 1,888 | +20% | 0 | 0 | — |
case-13 | fail→pass | 3,890 | 31,801 | +718% | 1 | 1 | 0% | 624 | 1,835 | +194% | 0 | 0 | — |
case-14 | fail→pass | 3,603 | 2,219 | -38% | 1 | 1 | 0% | 590 | 1,993 | +238% | 0 | 0 | — |
case-15 | fail→pass | 4,232 | 2,757 | -35% | 1 | 1 | 0% | 653 | 1,877 | +187% | 0 | 0 | — |
case-16 | pass→pass | 10,589 | 2,409 | -77% | 1 | 1 | 0% | 1,867 | 1,969 | +5% | 0 | 0 | — |
case-17 | pass→pass | 4,505 | 31,725 | +604% | 1 | 1 | 0% | 632 | 1,794 | +184% | 0 | 0 | — |
case-18 | fail→pass | 9,401 | 2,196 | -77% | 1 | 1 | 0% | 1,492 | 1,909 | +28% | 0 | 0 | — |
case-19 | pass→pass | 10,864 | 1,783 | -84% | 1 | 1 | 0% | 1,898 | 1,825 | -4% | 0 | 0 | — |
case-20 | pass→pass | 5,252 | 4,135 | -21% | 1 | 1 | 0% | 1,035 | 2,372 | +129% | 0 | 0 | — |
case-21 | pass→pass | 9,870 | 40,091 | +306% | 1 | 1 | 0% | 1,960 | 3,597 | +84% | 0 | 0 | — |
case-22 | pass→pass | 7,692 | 7,260 | -6% | 1 | 1 | 0% | 1,540 | 2,989 | +94% | 0 | 0 | — |
DecimalAI ran this skill against gemini-3.6-flash twice over the same eval suite — once with the skill loaded and once without — and compared the two runs case by case. 22 cases were attempted. The headline lift of +55 percentage points is the difference between those two pass rates over the 22 comparable cases.
Without the skill loaded, the model failed this case. With it loaded, the same prompt on the same model passed. This is one improved case from the latest verified run; every case, including any that regressed, is in the table above.
| Model | Method | Date | Lift |
|---|---|---|---|
| gemini-3.6-flash | verified | 8/13/2026 | +43% |
Other measured skills in the registry, with their headline benchmark lift.