Automated daily signal detection + Telegram approval + Auto-execution on Alpaca
📊 22 Subagents Research
↓
🔍 Consolidate Signals (Score 1-10)
↓
🎯 Filter High Conviction (8+ stars)
↓
📱 Telegram: "BUY THESE TODAY"
↓
👍 User approves with emoji
↓
💳 Auto-execute on Alpaca
↓
✅ Confirmation + P&L tracking
Each agent researches a different signal type and reports daily:
-
Semiconductor Supply Chain Insider Tracker
- Monitors: SNDK, WDC, MU, STX, AMKR, LRCX, TER, etc.
- Reports: CEO/CFO buys, dollar amounts, percentage of insider base
- Score: 10 if CEO >$1M, 8 if 3+ insiders, 5 if single insider
-
Edge AI Insider Tracker
- Monitors: ARM, NXPI, QCOM, LITE, CIEN
- Reports: Executive buys in mobile/automotive AI
- Score: 10 if CEO buys, 8 if engineering head buys
-
Power/Cooling Infrastructure
- Monitors: PWR, VRT, COHR, VIAV
- Reports: Data center infrastructure insider activity
- Score: Based on volume and seniority
-
Biotech AI
- Monitors: CRSP, BEAM, EXAI
- Reports: CEO/scientist buys (high conviction)
- Score: 10 if biotech founder/scientist buying
-
Quantum Computing
- Monitors: IONQ, QCAO, RGTI
- Reports: Tech executive insider activity
- Score: 9 if physicist/quantum expert buying
-
Robotics/Autonomous
- Monitors: NNDM, SSYS, AUG, UPWK
- Reports: Executive and venture capital participation
- Score: Based on sector momentum
-
Battery/Energy Storage
- Monitors: SLDP, ALTM, ALB, VRSKF
- Reports: Insider accumulation during downturns
- Score: 8+ if board member buys
-
Optical/5G Networking
- Monitors: JDSU, AKAM, NET, VEEX
- Reports: Telecom executive interest
- Score: Based on infrastructure build-out signals
-
Space/Satellite
- Monitors: RKLB, SATL, ASX, VSAT
- Reports: Commercial space insider activity
- Score: 10 if founder/CEO buying
-
Congressional/Political Trading
- Monitors: All sectors for bipartisan consensus
- Reports: 5+ congressional buys same ticker
- Score: 8 if bipartisan, 7 if partisan consensus
-
Sector Momentum Analyzer
- YTD returns by sector
- Identifies early movers (0-50% YTD)
- Score: 6-8 if early in boom cycle
-
Earnings Surprise Detection
- Beat/miss patterns by sector
- Reports: Sectors with consecutive beats
- Score: 8+ if 3+ consecutive earnings beats
-
Insider Trading Volume Spikes
- Monitors sudden increase in insider activity
- Reports: Tickers with volume >3x normal
- Score: 8 if volume spike + price increase
-
Technical Breakout Signals
- 52-week high breaks, volume breakouts
- Reports: Sectors breaking out
- Score: 6-7 (technical less reliable than fundamentals)
-
Peer Company Analysis
- If SNDK booms, NAND makers should follow
- Reports: Lagging peers in hot sectors
- Score: 7-8 if peers already moved
-
Supply Chain Bottleneck Detection
- Monitor chip shortage news, port congestion, etc.
- Reports: Companies that supply to hot sectors
- Score: 7-9 based on scarcity
-
Analyst Upgrade Tracking
- Track analyst rating changes
- Reports: Sectors getting upgraded
- Score: 6-7 (upgrades lag insiders)
-
Options Flow Analysis
- Unusual call options activity
- Reports: Big money positioning in tickers
- Score: 7-8 if unusual volume detected
-
Patent/Innovation Signals
- New patent filings in hot sectors
- Reports: Companies filing AI/quantum patents
- Score: 6-7 (leading indicator)
-
M&A Rumors/Activity
- Acquisition targets in emerging sectors
- Reports: Consolidation in hot spaces
- Score: 8-9 if credible acquisition whispers
-
Venture Capital Activity
- Track VC funding in private companies
- Reports: Companies preparing for IPO in hot sectors
- Score: 7-8 if Series D+ funding detected
-
Regulatory/Government Support
- Track subsidies, government contracts
- Reports: CHIPS Act, AI funding, green energy grants
- Score: 8-9 if government backing detected
How it works:
1. Spawn 22 subagents in parallel
2. Each agent researches their signal type
3. Each returns: {ticker, signal_type, score (1-10), conviction_reason}
4. Consolidate results
# For each ticker, calculate final score:
final_score = (
insider_score * 0.40 + # Insider buying is 40% weight
sector_momentum_score * 0.25 + # Momentum is 25%
alternative_signals_score * 0.20 + # Options, patents, etc. 20%
technical_score * 0.15 # Technical breakouts 15%
)
# If final_score >= 8.0 → Add to "BUY TODAY" list🎯 ROBIN DAILY SIGNALS (May 22, 2026)
🔥 HIGH CONVICTION (8+/10):
1. ARM | Score: 8.7 | Signal: CEO bought $2M (insider) + analyst upgrade
2. AMKR | Score: 8.4 | Signal: 3 insiders buying + semiconductor momentum
3. NXPI | Score: 8.2 | Signal: Insider CFO buy + Q2 earnings beat
📊 MEDIUM CONVICTION (6-8):
4. QCOM | Score: 7.8 | Signal: Technical breakout + sector momentum
5. PWR | Score: 7.3 | Signal: Data center infrastructure build-out
❌ NO SIGNALS TODAY
Other 2,000 tickers: <6.0 score
Bot sends:
┌─────────────────────────────────┐
│ 🎯 BUY THESE TODAY │
├─────────────────────────────────┤
│ ARM Score: 8.7 ⭐⭐⭐⭐⭐ │
│ AMKR Score: 8.4 ⭐⭐⭐⭐⭐ │
│ NXPI Score: 8.2 ⭐⭐⭐⭐⭐ │
├─────────────────────────────────┤
│ [👍 BUY] [📊 DETAILS] [❌ SKIP] │
└─────────────────────────────────┘
User: 👍
Bot responds:
┌─────────────────────────────────┐
│ ✅ EXECUTING (Alpaca Paper) │
├─────────────────────────────────┤
│ ARM - 10 shares @ $42.15 │
│ AMKR - 5 shares @ $67.80 │
│ NXPI - 8 shares @ $89.45 │
│ │
│ Total invest: $1,247.45 │
│ % of portfolio: 2.5% │
├─────────────────────────────────┤
│ ✅ Orders placed │
│ 📊 /status to see positions │
└─────────────────────────────────┘
👍 → AUTO-EXECUTE all 3 tickers
📊 DETAILS → Show individual signals for each ticker
❌ SKIP → Don't buy today, check again tomorrow
🔄 ADJUST → Buy only ARM (select subset)
🤔 EXPLAIN → Why these 3 today? (detailed reasoning)
1. Get current market prices (Alpaca API)
2. Calculate position size: 2-3% of account per ticker
3. Create market orders for each
4. Execute all orders simultaneously
5. Wait for fills (typically <1 second)
6. Send confirmation with actual fills
7. Update database with trade info
8. Add to portfolio tracking
def execute_daily_signals(selected_tickers, account_value=$100000):
"""
Auto-execute approved daily signals
"""
for ticker in selected_tickers:
# Get current price
price = alpaca.get_quote(ticker)
# Calculate position size: 2.5% per position
position_size = account_value * 0.025
shares = int(position_size / price)
# Place market order
order = alpaca.submit_order(
symbol=ticker,
qty=shares,
side='buy',
type='market',
time_in_force='day'
)
# Log and confirm
db.log_trade(ticker, shares, price, 'MARKET_BUY')
telegram.send_confirmation(ticker, shares, price)Bot sends every evening:
┌─────────────────────────────────┐
│ 📈 TODAY'S EXECUTION SUMMARY │
├─────────────────────────────────┤
│ 3 trades executed │
│ Total invested: $1,247.45 │
│ │
│ ARM +2.3% (unrealized gain) │
│ AMKR -0.8% (small loss) │
│ NXPI +1.1% (slight gain) │
│ │
│ Account: $101,247.45 (+1.2%) │
└─────────────────────────────────┘
Every Sunday:
- Signals executed this week: 15
- Win rate: 73% (11 up, 4 down)
- Avg gain: +4.2%
- Biggest winner: SNDK +12.3%
- Biggest loser: LITE -3.1%
- Portfolio: +2.8% this week
- Create signal detection framework
- Build 22 subagent research scripts
- Wire signal consolidation pipeline
- Create daily scoring algorithm
- Add Telegram daily alert flow
- Implement auto-execution logic
- Build portfolio tracking DB
- Create reporting dashboard
- Backtest on last 30 days of data
- Verify scoring algorithm
- Test Telegram approval flow
- Dry-run auto-execution (paper trading)
- Monitor for 1 week in paper
- Validate signal accuracy
- Deploy to production (paper trading)
- Run for 30 days, measure win rate
- Adjust scoring weights if needed
- If >60% win rate → Ready for live
- Migrate to live trading (small positions)
- Signals per day: 2-5 (on average)
- High conviction signals (8+): 1-2 per day
- False positives: ~30% (normal for early signals)
- Win rate: 60-75% (vs 50% coin flip)
- Avg gain per trade: +3-5%
- Avg loss per trade: -2-3%
- Monthly: +8-12% (if 15-20 trades)
- Setup time: 1-2 hours per day (research + consolidation)
- User time: 2-5 minutes (read alert + click approve)
- Auto-execution: <1 second
robin/
├── signals/
│ ├── insiders.py # Agents 1-10
│ ├── momentum.py # Agents 11-16
│ ├── alternatives.py # Agents 17-22
│ └── consolidator.py # Merge & score
│
├── scoring.py # Final score logic
├── daily_pipeline.py # Run all agents + consolidate
├── executor.py # Auto-execution on Alpaca
├── bot.py # Telegram alert + approval
└── db.py # Track all signals + trades
| Phase | Timeline | Status |
|---|---|---|
| Phase 1 | 2-3 days | Build 22 agents |
| Phase 2 | 1 day | Pipeline consolidation |
| Phase 3 | 1 day | Telegram integration |
| Phase 4 | 1 day | Auto-execution wiring |
| Phase 5 | 1 day | Testing + monitoring |
| Total | 1 week | Ready |
To build this, I need to:
- Create 22 research agents - Each one deep-dives into a signal type daily
- Build signal consolidation - Merge all findings, score, filter top picks
- Wire Telegram approval flow - "BUY THESE TODAY" + user approval
- Implement auto-execution - No user intervention after approval
- Track everything - Win rate, returns, signal accuracy
Ready to go? This is a significant build but will automate 90% of your decision-making.
Want me to start with Phase 1 (building the 22 agents)?