Skills for building, testing, and deploying trading strategies using Robonet's MCP server.
This repository provides 7 focused skills for quantitative trading with Robonet, organized by development lifecycle:
1. browse-robonet-data → Explore available resources (symbols, indicators, strategies)
2. design-trading-strategies → AI-powered strategy ideation ($0.05-$1.00)
3. build-trading-strategies → Generate complete strategy code ($1-$4.50)
4. improve-trading-strategies → Refine, optimize, add ML predictions ($0.50-$4)
5. test-trading-strategies → Backtest and validate performance ($0.001)
6. deploy-live-trading → Deploy to live trading (⚠️ HIGH RISK, $0.50)
7. trade-prediction-markets → Polymarket YES/NO token trading
browse-robonet-data
→ design-trading-strategies (explore ideas cheaply)
→ build-trading-strategies (create code)
→ test-trading-strategies (validate)
browse-robonet-data
→ build-trading-strategies (create code)
→ improve-trading-strategies (refine iteratively)
→ test-trading-strategies (validate each improvement)
→ deploy-live-trading (⚠️ only after thorough testing)
browse-robonet-data
→ trade-prediction-markets (self-contained Polymarket workflow)
Purpose: Fast, low-cost exploration of trading resources Tools: 8 data tools (get_all_symbols, get_all_indicators, etc.) Cost: Free to $0.001 Risk: None (read-only) Use when: Starting any workflow, checking data availability
Purpose: AI-powered strategy ideation Tools: 1 tool (generate_ideas) Cost: $0.05-$1.00 per generation Risk: Low (ideas only, no execution) Use when: Exploring concepts before expensive development
Purpose: Generate complete strategy code Tools: 2 tools (create_strategy, create_prediction_market_strategy) Cost: $1.00-$4.50 per strategy (MOST EXPENSIVE) Risk: Low (generates code but doesn't execute) Use when: Ready to commit to development with clear requirements
Purpose: Refine, optimize, and enhance existing strategies Tools: 3 tools (refine_strategy, optimize_strategy, enhance_with_allora) Cost: $0.50-$4.00 per operation Risk: Low (modifies code but doesn't execute) Use when: Have existing strategy that needs improvement
Purpose: Validate strategy performance on historical data Tools: 3 tools (run_backtest, run_prediction_market_backtest, get_latest_backtest_results) Cost: $0.001 per backtest Risk: None (simulation only) Use when: After building or improving strategies (ALWAYS before deployment)
Purpose: Deploy strategies to live trading on Hyperliquid
Tools: 5 tools (deployment_create, list, start, stop, account management)
Cost: $0.50 to create deployment
Risk:
CRITICAL WARNING: NEVER deploy without extensive backtesting. You can lose ALL deployed capital.
Purpose: Build and test Polymarket prediction market strategies Tools: 6 tools (event browsing, data analysis, strategy creation, backtesting) Cost: $0.001 for data, $1-$4.50 for strategy creation Risk: Medium (specialized domain) Use when: Trading on real-world events (politics, economics, sports, crypto)
Add skills using the Agent Skills CLI:
# Install specific skills
npx skills add https://github.com/robonet-tech/skills --skill browse-robonet-data
npx skills add https://github.com/robonet-tech/skills --skill build-trading-strategies
npx skills add https://github.com/robonet-tech/skills --skill test-trading-strategies
# Or install all skills
npx skills add https://github.com/robonet-tech/skills --skill browse-robonet-data
npx skills add https://github.com/robonet-tech/skills --skill design-trading-strategies
npx skills add https://github.com/robonet-tech/skills --skill build-trading-strategies
npx skills add https://github.com/robonet-tech/skills --skill improve-trading-strategies
npx skills add https://github.com/robonet-tech/skills --skill test-trading-strategies
npx skills add https://github.com/robonet-tech/skills --skill deploy-live-trading
npx skills add https://github.com/robonet-tech/skills --skill trade-prediction-markets- None: browse-robonet-data, test-trading-strategies (read-only/simulation)
- Low: design-trading-strategies, build-trading-strategies, improve-trading-strategies (code generation)
- Medium: trade-prediction-markets (specialized domain)
- HIGH: deploy-live-trading (real capital at risk)
- Free/$0.001: Browse data, backtest (use liberally)
- $0.05-$1.00: Generate ideas (cheap exploration)
- $0.50-$4.50: Build/improve strategies (expensive AI operations)
- $0.50: Deploy to live trading (setup fee)
- Data Analyst: browse-robonet-data
- Strategy Researcher: browse-robonet-data, design-trading-strategies
- Strategy Developer: build-trading-strategies, improve-trading-strategies, test-trading-strategies
- Live Trader: test-trading-strategies, deploy-live-trading
- Prediction Market Trader: trade-prediction-markets
Each skill provides contextual information beyond tool definitions:
- Purpose & Context: WHY the skill exists, what problems it solves
- Decision Logic: WHEN to use specific tools (explicit guidance)
- Workflows: HOW to sequence tools for complex goals (step-by-step)
- Domain Knowledge: WHAT "good" looks like (metrics, best practices, trading expertise)
- Safety Guardrails: Risk management, constraints, troubleshooting patterns
- Cost/Time Awareness: Help LLM make intelligent trade-offs
This transforms the LLM from "a tool caller" into "a domain expert who uses tools effectively."
| Skill | Tools | Execution Time | Cost | Risk |
|---|---|---|---|---|
| browse-robonet-data | 8 | <1s | Free-$0.001 | None |
| design-trading-strategies | 1 | 20-40s | $0.05-$1.00 | Low |
| build-trading-strategies | 2 | 30-60s | $1.00-$4.50 | Low |
| improve-trading-strategies | 3 | 20-60s | $0.50-$4.00 | Low |
| test-trading-strategies | 3 | 20-60s | $0.001 | None |
| deploy-live-trading | 5 | <5s | $0.50 setup | HIGH |
| trade-prediction-markets | 6 | Variable | $0.001-$4.50 | Medium |
Total: 28 tool references across 7 skills (some tools shared between skills)
Common documentation for all skills:
- tool-catalog.md: Complete reference for all 24 MCP tools
- Additional shared references to be added for performance metrics, risk management, etc.
To contribute new skills or improve existing ones, please follow the Agent Skills format.
- Issues: https://github.com/robonet-tech/skills/issues
- Robonet Dashboard: https://robonet.tech
- Documentation: See individual skill SKILL.md files
[License information to be added]