CreatorSamin Yasar
Sourceyoutu.be/lH5wrfNwL3k
PublishedApril 6, 2026
Transcript Date2026-06-07
Views1,432,704
Creator BioEx-JP Morgan, Ex-Amazon software engineer; CS degree NYU; AI automation consultant

Summary

Samin Yasar, a former JP Morgan and Amazon software engineer, walks through three levels of using Claude to automate stock market activity using the Alpaca paper trading API. The core argument is that Wall Street's advantage over retail traders comes down to three things -- data access, execution speed, and analytical intelligence -- and that Claude with the right integrations can close all three gaps simultaneously.

Level one covers connecting Claude Desktop to Alpaca's paper trading API to place test trades by voice. Level two builds on that with an automated trailing stop and ladder-buy bot that runs on a cron schedule during market hours. Level three covers options trading mechanics and the wheel strategy, where Claude monitors positions, selects strike prices, and rolls contracts on a schedule.

A copy-trading module is also demonstrated, using Capitol Trades (a public database of congressional stock disclosures) plugged into Claude via MCP to automatically mirror the trades of consistently outperforming politicians. The video uses paper trading throughout and includes a disclaimer that none of this constitutes financial advice.

Key Points

Quotable Moments

Quotable moments are auto-generated from the transcript. Speaker attribution and quote accuracy should be verified against the original source before republishing or sharing.

Samin Yasar -- on the data gap
"Imagine a poker game, and you're sitting at a table with your two cards trying to figure out what to do. But the person across from you can see every card on the table and every card in your hand."
Why it works: Makes an abstract structural disadvantage visceral. The poker metaphor collapses a complex market dynamic into a single image that most people can feel immediately.
Samin Yasar -- on encoding strategy
"The rules aren't the limitation. That's the whole point. This is how you take what you know -- your instincts, your risk tolerance, your read on the market -- and you encode it."
Why it works: Reframes AI automation not as replacing human judgment but as amplifying it. Addresses the common fear that automated trading removes the trader from the equation.
Samin Yasar -- on selling options
"The insurance companies make billions doing this. They collect premiums after premiums from millions of people and pay out on a small percentage of claims. The math basically works in their favor over time."
Why it works: Reframes options selling from a complex financial instrument into a business model most people already intuitively understand. Reduces friction for new learners.
Samin Yasar -- on the wheel strategy outcome
"You got Tesla cheaper than anyone else, and you wanted it anyways... That's not a loss, that's your plan working."
Why it works: Reframes assignment (being forced to buy the stock) as the intended outcome of a well-designed strategy rather than a failure. Shifts the reader's mental model of risk.

Concepts and Ideas

Core Framework -- The Three Gaps

The Data Gap

Institutional traders and politicians have access to real-time flows of information that retail investors see only after the fact -- if at all. The presenter frames this as the fundamental structural disadvantage retail traders face. Data services that aggregate congressional disclosures, options flow, and whale activity compress this gap significantly.

The Execution Gap

Wall Street operates on automated systems that monitor positions and act around the clock. A retail trader who spots a good setup but steps away for 10 minutes may miss the window entirely. Claude connected to a brokerage API closes this gap by running rules-based execution on a schedule.

The Intelligence Gap

Having data and speed means nothing without a coherent plan. Institutional desks have teams of analysts who interpret information and build structured strategies. AI assistants like Claude can now perform much of this function -- reading data, applying rules, and acting -- at no marginal cost once configured.

Trading Mechanics

Trailing Stop Loss

A trailing stop loss is a dynamic floor price set a percentage below the current stock price. As the stock rises, the floor rises with it. If the stock falls and hits the floor, the position is sold automatically. The floor never moves down, only up -- locking in gains at every step while capping the maximum loss to the initial gap.

Ladder Buys

Rather than buying all shares at once, ladder buys pre-schedule additional purchases at incrementally lower prices. If a stock drops 15%, you buy more. If it drops 30%, you buy more again. This systematically lowers your average cost basis during a downturn rather than sitting on a growing loss.

Paper Trading

Paper trading uses a simulated account funded with fake money but connected to real market prices and real order book mechanics. It allows a trader to test and validate strategies under live conditions without any actual capital at risk. Alpaca provides a paper trading environment through its API.

Options -- Calls and Puts

A call option gives the buyer the right to purchase a stock at a locked-in price (strike price) before expiration. A put option gives the buyer the right to sell at a locked-in price. Both are contracts between two parties: the buyer pays a premium for the right; the seller collects the premium and takes on the obligation.

Selling Options -- Becoming the House

When you sell an option rather than buy one, you flip the table and collect the premium upfront. Your profit comes when the contract expires worthless -- the other party never exercises it. This mirrors how insurance companies profit: collect premiums from many, pay out on a fraction of claims. The statistical edge favors the seller over time.

Advanced Strategies

The Wheel Strategy

The wheel is a two-stage cyclical options income strategy. Stage one: sell a cash-secured put below the current price. Collect the premium. If the option expires worthless, repeat. If assigned, acquire shares at a discount. Stage two: sell a covered call above your cost basis on those shares. Collect more premium. If the shares get called away, go back to stage one. The wheel generates income regardless of whether the underlying stock trends up, down, or sideways.

Copy Trading via Public Disclosures

Members of US Congress are legally required to disclose their stock trades within 45 days under the STOCK Act. Services like Capitol Trades aggregate and publish this data. Because congressional members often trade on advance knowledge of policy, regulation, or government contracts, their portfolios historically outperform the broader market. Mirroring those trades is a legal, data-driven strategy available to any retail investor.

Technical Infrastructure

MCP (Model Context Protocol)

MCP is the protocol that allows Claude to connect to external data services and APIs as live tools rather than static knowledge. The presenter uses the analogy of a power outlet: data services run information through their systems, and MCP is the plug that lets Claude tap into that feed in real time. Once connected, Claude can query those services dynamically as part of any conversation or scheduled task.

Cron Job Scheduling

A cron job is a scheduled task that runs automatically at set intervals on a computer. In this context, Claude uses cron scheduling (via Claude Code) to check positions, evaluate trailing stop levels, and execute trades during market hours without manual intervention. The schedule is set once and runs as long as the machine is on.

Brokerage API Access

Traditional brokerages require a human to place trades. Modern platforms like Alpaca expose an API that lets external software -- including Claude -- authenticate with API keys and execute orders programmatically. This is the technical bridge that allows a language model to interact with real (or simulated) financial markets.

Implementation

Implementation steps are auto-generated from the transcript content and are provided for informational purposes only. They do not constitute professional advice of any kind. Always consult a qualified professional before acting on any information presented here.

1

Download the Claude Desktop App

Go to claude.com/download and install the Claude Desktop app for Mac or Windows. Make sure your operating system is up to date -- the desktop app requires a reasonably recent OS version to support all features. You will need either a Claude Pro or Claude Max subscription to access the full capability set used in this workflow, including the Claude Code interface and scheduling functions. Sign in and confirm the app launches correctly before proceeding.

2

Create a Free Alpaca Account

Go to alpaca.markets and create a free account. You do not need to fund a live account -- the paper trading environment is what you want. After signing up and completing the basic verification steps, open the Alpaca dashboard and navigate to the paper trading section. Create a new paper trading account (the video uses a $50,000 starting balance as a working example). Paper trading gives you real market prices and real order mechanics with zero capital at risk.

3

Generate Your Alpaca API Keys

Inside your Alpaca paper trading account, scroll to the API Keys section and click "Generate New API Keys." You will receive three values: an endpoint URL, an API key, and a secret key. Copy all three -- you will need them in the next step. Store them somewhere accessible but private; treat the secret key the same way you would treat a password. Do not commit these to any public repository or paste them into shared documents.

4

Connect Claude to Alpaca and Test a Trade

Open the Claude Desktop app and switch to the Code interface (the button labeled "code" in the interface). Create a new project folder (the video uses a folder named "trading" inside Documents to keep things organized). Paste your Alpaca endpoint, API key, and secret into the chat and tell Claude to save those credentials to a file in that folder so you do not have to re-enter them each session. Then send a simple test command -- something like "buy one share of Apple using my Alpaca paper trading account" -- and confirm the trade appears in your Alpaca dashboard. This verifies the connection is live before building anything more complex.

5

Define Your Trailing Stop Strategy in Plain English

Write out your trading rules conversationally and give them to Claude. At minimum, specify: which stock to trade, how many shares to start with, your stop loss threshold (e.g., sell if the stock drops 10% from purchase), your trailing stop trigger (e.g., if it climbs 10%, raise the floor to 5% below the new price), and any ladder buy rules (e.g., buy 10 more shares if it drops 20%, buy 20 more if it drops 30%). The more specific you are, the more precisely Claude can encode the rules. You can refine these over time -- the rules are not permanent, and Claude will follow updated instructions.

6

Schedule Claude to Monitor Positions Automatically

After your initial trade is placed and your rules are set, tell Claude to create a recurring schedule using the /schedule command. In the video, Claude is instructed to check Tesla positions every 5 minutes during market hours (Monday through Friday, 9:00 a.m. to 4:00 p.m. Eastern). Claude will create a cron job visible in the scheduled tasks panel on the left sidebar of Claude Code. As long as your computer is running and Claude is open, the bot will execute without any further input. Confirm the schedule appears in the task panel before stepping away.

7

Scenario-Test Your Bot Before Relying On It

Before treating any automated strategy as reliable, walk through hypothetical scenarios with Claude to verify it will behave the way you expect. Ask things like "What would you do if Tesla shot up to $500?" or "What happens if the stock gaps down 25% at the open?" Claude will explain its intended actions based on your encoded rules. If the responses reveal gaps or misalignments, adjust your instructions and re-test. This is a low-cost way to validate logic before it encounters real market conditions.

8

Set Up the Congressional Copy Trading Bot (Optional)

Create a separate paper trading account in Alpaca for this strategy so you can track its performance independently. Generate a new set of API keys for that account and start a new Claude session. Paste the new credentials and the URL for Capitol Trades (capitoltrades.com). Instruct Claude to identify a currently active congressional trader with a strong recent record, monitor their disclosed trades via Capitol Trades, and mirror those trades in your Alpaca paper account on a scheduled basis. Claude will handle the politician selection and explain its reasoning. You can ask it to explain its choice and backtest the strategy before letting it run live (in paper mode).

9

Build the Wheel Strategy Bot

The wheel strategy requires more precise instructions than the trailing stop bot because options have additional parameters -- strike price, expiration date, and contract type. Give Claude a full ruleset covering both stages. Stage one: sell cash-secured puts at a strike roughly 10% below current price, two to four weeks out. If the put expires worthless, sell again. If assigned, move to stage two. Stage two: sell covered calls at a strike roughly 10% above your effective cost basis, same expiration window. Never sell a call below cost basis. Close any contract early if it hits 50% profit. Check positions every 15 minutes during market hours and send a daily summary at market close. Claude handles all rolling decisions and execution against those rules.

10

Run in Paper Mode for an Extended Period Before Going Live

None of the strategies in this video were demonstrated with real money, and for good reason: automated trading bots behave differently in edge cases (earnings announcements, halted stocks, gap opens, low liquidity) than they do in typical conditions. Run your bots in paper mode for at least several weeks across different market conditions before considering any real capital. Track the results, review what the bot did at each decision point, and refine your rules. The goal of the paper period is not just profit confirmation -- it is learning how the rules behave so you can improve them with confidence before anything real is at stake.

Resources

Tools, platforms, and references mentioned in the video or linked in the description. Verify current availability and pricing before relying on any of these.

Core Setup -- Required

Claude Desktop App

The desktop application for Claude. Required for this workflow -- the web interface does not support the Claude Code environment or cron scheduling used to run automated bots. Available for Mac and Windows. A Pro or Max subscription is recommended for extended sessions and full feature access.

claude.com/download
RequiredFree DownloadSubscription Needed

Alpaca Markets

Commission-free brokerage platform with full API access and a built-in paper trading environment. The API is the bridge that allows Claude to place trades programmatically. Paper trading accounts are free to create and come with simulated capital. Real trading requires account verification and funding. Supports stocks and options.

alpaca.markets
RequiredFree to StartAPI AccessPaper Trading
Data -- Congressional Trading

Capitol Trades

Free public database aggregating stock trade disclosures filed by US members of Congress under the STOCK Act. Tracks which politicians are buying or selling which securities, when the trades were made, and when they were disclosed. The presenter uses this as the data feed for the copy trading bot, pointing Claude at it to identify active traders with strong recent performance.

capitoltrades.com
FreePublic DataCongressional Disclosures
Creator Resources

Claude Club (Skool Community)

The presenter's paid membership community on Skool. Includes the full prompt library referenced in the video (trailing stop, wheel strategy, copy trading bot prompts), along with deeper tutorials, live sessions, and community discussion. Mentioned as the place to go for the exact prompts demonstrated in the video.

skool.com/claude
Paid CommunityPrompt Library

AI Answers Resource Hub

The presenter's free resource hub on Skool. A secondary entry point to his educational materials, case studies, and AI automation guides for business owners.

skool.com/aianswers
FreeAI Automation
Underlying Technology

MCP (Model Context Protocol)

The open protocol developed by Anthropic that allows Claude to connect to external data services, APIs, and tools as live integrations rather than static knowledge. In this video, MCP is the mechanism that allows Claude to query Capitol Trades data in real time rather than relying on its training data. MCP servers can be configured for a wide range of services -- financial data providers, CRMs, productivity tools, and more. Documentation is available through Anthropic.

docs.anthropic.com -- MCP Overview
TechnicalOpen ProtocolFree

Alpaca Markets API Documentation

Technical reference for the Alpaca trading API. Covers authentication, order types, paper vs. live environments, options endpoints, and streaming data. Relevant for anyone who wants to understand what Claude is actually doing under the hood when it places orders, or who wants to extend beyond the conversational interface.

docs.alpaca.markets
TechnicalFree

Quiver Quantitative

Aggregates congressional trading disclosures alongside lobbying data, government contracts, and patent filings -- providing context beyond the raw trade data that Capitol Trades shows. Free tier includes congressional trade screening with sector tagging and member filtering. Useful for identifying cluster signals and committee-aligned trades across multiple data sources simultaneously.

quiverquant.com
Free TierCongressional DataLobbying + Contracts
Regulatory and Legal Context

STOCK Act (Stop Trading on Congressional Knowledge Act)

The 2012 US federal law that requires members of Congress, the President, Vice President, and certain executive branch officials to publicly disclose stock trades within 45 days of execution. This is the legal foundation that makes congressional trade data publicly accessible and is why services like Capitol Trades exist. Disclosures are filed with the Clerk of the House or the Secretary of the Senate and are posted publicly.

House Financial Disclosures -- EFTS Search
Legal ContextPublic Record
Note: The presenter emphasizes throughout that all demonstrations use paper trading (simulated accounts with fake money). Nothing in this video constitutes financial advice. Verify the current terms, fees, and capabilities of any platform before using it. Options trading in particular carries substantial risk and may not be available in all account types or jurisdictions.

IBKR Setup for Canadian Traders Bonus

These steps reflect general setup guidance for Interactive Brokers as of mid-2026. IBKR updates its platform regularly and specific UI steps or settings may drift. Verify current platform behaviour and documentation at ibkr.ca before relying on any specific step. Nothing here constitutes financial or investment advice.

Interactive Brokers is the practical path for Canadian residents who want to move beyond Alpaca's paper trading environment. IBKR is registered with IIROC, supports both stocks and options, offers a full API, and has a paper trading environment that mirrors the live platform. The setup is more involved than Alpaca's key-and-secret model but is well-documented and stable once configured.

1

Open an IBKR Account

Go to ibkr.ca and open an individual account. You will need to complete identity verification -- have government ID and a recent proof of address ready. During the application, apply for options trading permissions at Level 2 minimum (covers covered calls and cash-secured puts, which are required for the wheel strategy). If you intend to trade spreads or LEAPS, apply for Level 3. Approval typically takes one to three business days. Once your live account is approved, IBKR will send separate credentials for the paper trading environment.

2

Choose Between TWS and IB Gateway

IBKR offers two applications that serve as the local API bridge. Trader Workstation (TWS) is the full-featured trading platform with charts, order management, and a complete UI. IB Gateway is a stripped-down application with no trading interface -- it exists solely to provide API access. For a dedicated always-on machine running automated bots, IB Gateway is the right choice: it uses significantly less memory, starts faster, and is designed for exactly this use case. Download IB Gateway from the IBKR website under Traders > Trading Software.

3

Enable the API and Configure Connection Settings

This is the step where most people stall. In TWS, go to Edit > Global Configuration > API > Settings. In IB Gateway, the same settings are under Configure > Settings > API > Settings. Enable "Enable ActiveX and Socket Clients." Set the socket port -- note these defaults carefully: TWS live uses 7497, TWS paper uses 7496, IB Gateway live uses 4001, IB Gateway paper uses 4002. Connecting to the wrong port produces a connection refused error with no helpful explanation. Check "Allow connections from localhost only" for security. Uncheck "Read-Only API" -- the bot needs to place orders, not just read data. Save and restart the application for settings to take effect.

4

Disable Auto-Logoff

By default, TWS and IB Gateway log themselves out at a set time each day -- often around midnight Eastern. Do this before anything else. In TWS: Edit > Global Configuration > Lock and Exit > Auto logoff timer -- disable it or set it to a time outside your trading window. In IB Gateway: Configure > Settings > Lock and Exit. If you skip this step, your bot will go dark every night without warning and you will have open positions with no monitoring. This is one of the most common causes of silent bot failure on IBKR.

5

Install ib_insync and Test the Connection

Tell Claude to set up a Python environment in your trading folder and install the ib_insync library: pip install ib_insync. This is a well-maintained third-party wrapper around IBKR's official ibapi library that makes API interactions significantly more readable and practical. IBKR's official ibapi works but is verbose and callback-heavy -- ib_insync is the standard choice for this kind of bot work. Ask Claude to write a simple connection test: connect to IB Gateway on localhost at your paper port (4002), request your account summary, and print the result. If your paper account balance comes back, the connection is confirmed.

6

Manage Client IDs

Every connection to the IBKR API requires a unique integer client ID. If two scripts connect simultaneously with the same client ID, one is dropped without a clear error message. Assign a distinct client ID to each bot or script -- for example, 1 for the trailing stop bot, 2 for the wheel strategy bot, 3 for the congressional copy trading bot. Tell Claude to hardcode these IDs into each script and never reuse them across simultaneous connections. If you see a connection that drops immediately after establishing, a client ID conflict is the most likely cause.

7

Choose Your Always-On Machine

IB Gateway must be running for any scheduled bot to execute. This means the machine running it needs to be on, connected, and stable during trading hours. A dedicated low-cost always-on computer is the practical solution -- a basic mini PC (Intel NUC, Beelink, or similar) running Windows or Linux costs $150-300 CAD and draws very little power. A spare laptop works equally well. The key requirement is that it stays on and connected during market hours without depending on your primary workstation. A Raspberry Pi can run IB Gateway on Linux with a virtual display configured, but that setup is covered in a separate guide. Remote access via Tailscale (free, minimal configuration, works through home routers without port forwarding) lets you check on the machine from anywhere.

8

Add a Watchdog Script

Ask Claude to write a simple watchdog script that runs on a schedule and checks whether IB Gateway is still alive. If the process is not running, the watchdog restarts it automatically. This is a one-time setup that meaningfully improves reliability -- IB Gateway occasionally crashes or freezes, especially after software updates. A watchdog running every 5 minutes eliminates most unplanned downtime without requiring manual intervention. Claude can generate this script in Python in under a minute; it belongs in the same trading folder as your other scripts.

9

Note on Canadian Currency Routing

Canadian residents trading US-listed equities through IBKR will encounter CAD/USD conversion. IBKR's automatic currency conversion carries a wider spread than their Ideal Pro forex mechanism. For meaningful position sizes, convert currency manually via Ideal Pro (under the currency conversion section in TWS or by placing a forex order through the API) before buying US equities or options. The difference is small per trade but accumulates over a full wheel strategy cycle. Tell Claude to flag when a trade would require currency conversion so you can handle it deliberately rather than letting it happen automatically at the wider spread.

Troubleshooting -- Steps 3 and 4

These are the specific errors most people hit during initial IBKR API configuration, with the exact fix for each.

Connection refused / ConnectionRefusedError on localhost
Most likely cause: wrong port number, or TWS/IB Gateway is not running. Confirm the application is open and logged in. Confirm you are using the correct port for your environment (paper vs. live, TWS vs. IB Gateway). Also check that "Enable ActiveX and Socket Clients" is checked in API settings and that you have restarted the application after saving settings.
Connection establishes then drops immediately
Most likely cause: client ID conflict. Another script or session is already connected with the same client ID integer. Change the client ID in your script to a different number and reconnect.
Error 507 -- FA profile is not assigned
Most likely cause: your account is set up as a financial advisor account rather than an individual account. Check your account type in the IBKR client portal. Individual accounts do not use FA profiles.
Firewall blocking localhost connection (Windows)
Most likely cause: Windows Firewall is blocking the socket connection even though you have enabled it in TWS/IB Gateway settings. Go to Windows Firewall > Advanced Settings > Inbound Rules and confirm that TWS or IB Gateway has an allow rule. Alternatively, temporarily disable the firewall to confirm this is the cause, then add the rule specifically rather than leaving the firewall off.
Orders rejected -- Not subscribed to requested market data
Most likely cause: your IBKR account does not have a market data subscription for the exchange you are routing to. For paper trading this is rarely a problem, but on a live account you may need to subscribe to US Options (OPRA) data. Paper trading accounts use delayed data by default -- add a line in your ib_insync connection to request delayed data explicitly if you are seeing this on paper: ib.reqMarketDataType(3).
Bot goes dark overnight / stops executing
Most likely cause: auto-logoff is enabled (see Step 4). Secondary cause: the machine went to sleep. Check power settings on your dedicated machine and set it to never sleep. Confirm the watchdog script from Step 8 is running so IB Gateway restarts if it crashes.
Options orders rejected -- insufficient permissions
Most likely cause: your account's options permission level does not cover the strategy you are attempting. Selling cash-secured puts requires Level 2. Spreads require Level 3. Log into the IBKR client portal, go to Account Settings > Trading Permissions, and apply for the required level. Approval is typically same-day for existing accounts.
Power failure risk: if your dedicated machine loses power while the bot has open positions, those positions remain in the market unmonitored until the machine restarts and the bot reconnects. A UPS (uninterruptible power supply) in the 600-900VA range -- approximately $100-150 CAD from APC or CyberPower -- provides enough runtime to survive short outages or complete a graceful shutdown. Given that a single unmonitored position can exceed the cost of the UPS many times over, this is a risk management decision, not an optional accessory.
IBKR paper trading accounts reset periodically. Positions and P&L may clear without warning. Do not interpret a reset as a bot error. Run paper testing long enough across multiple reset cycles to confirm your strategy logic survives the event cleanly.
IB Gateway requires occasional manual restarts after IBKR software updates, which are pushed roughly monthly. The watchdog script handles crashes but not update-prompted restarts that require user confirmation. Check your dedicated machine after any IBKR update notification.

Following Congressional Trades Bonus

This tab is provided for informational and educational purposes only. Nothing here constitutes financial or investment advice. Options trading involves substantial risk of loss. Congressional trade data is public and historical outperformance is an aggregate statistical observation -- it does not guarantee results on any individual trade. Always consult a qualified professional before acting on any information presented here.

Disclosure lag is the central constraint of this strategy. The STOCK Act requires congressional members to disclose trades within 45 days. Many file late. Average real-world disclosure lag is 20-30 days for compliant filers. Every strategy and tool recommendation on this tab is shaped by that reality. Strategies that require acting on fresh information do not fit this data source. Strategies with duration and defined risk do.

Calls and Puts

A call option gives you the right to buy a stock at a locked-in price (the strike price) before the contract expires. You buy calls when you expect the stock to rise. A put option gives you the right to sell a stock at a locked-in price. You buy puts when you expect the stock to fall -- or to protect a position you already own. Every option contract covers 100 shares.

Premium, Strike, and Expiry

Every option has three defining parameters. The premium is the price you pay (or collect) for the contract -- it decays toward zero as expiration approaches, which is called theta decay. The strike price is the price at which the contract can be exercised. The expiry date is the last day the contract is valid. Buying options costs premium and subjects you to time decay. Selling options collects premium and creates an obligation to the buyer.

Defined vs. Undefined Risk

A spread (buying one option and selling another at a different strike) caps both your maximum gain and your maximum loss -- this is defined risk. You know exactly what you can lose before you enter. An outright long call or long put has a maximum loss equal to the premium paid -- also defined, and generally the right approach for directional bets. A short put (selling a put without owning the shares) has a maximum loss equal to the strike price minus the premium collected -- substantial if the stock collapses. Understanding which category a strategy falls into before sizing a position is the most important mechanical concept on this tab.

Understanding the Signal

Congressional trade disclosures are public record. The STOCK Act requires members of Congress, the President, and Vice President to report stock transactions within 45 days of execution. The data covers transaction date, ticker, trade type (buy/sell/option), and a dollar range (exact amounts are not disclosed -- ranges like $15,001-$50,000 are typical). Services like Capitol Trades and Quiver Quantitative aggregate and present this data in filterable, searchable form.

The statistical case for following congressional trades is real but requires honest framing. Studies of aggregate congressional portfolios show consistent outperformance versus the S&P 500 -- the presenter cites a 34.8% return for one member versus 15% for the index over the same period. But this is an aggregate observation across many trades and members. Individual trades fail. Individual members have losing periods. You are using a probabilistic edge, not insider information -- and the edge is amplified or destroyed depending on how well you filter and size your positions.

Filtering for High-Quality Signals

Not all congressional trades carry equal weight. The following filters separate high-quality signals from noise.

Filter What to Look For Why It Matters
Cluster buying Multiple members buying the same ticker within a short window, especially across party lines Single trades may reflect personal conviction or an advisor's call. Cross-party cluster trades suggest a shared information environment that is harder to explain as coincidence
Committee alignment Member sits on a committee with oversight of the sector they are trading A member of Armed Services buying a defence contractor has structural information access. A random member doing the same trade carries less signal weight
Position size Large range ($100,000+) relative to the member's typical trading frequency A $15,000 purchase from an active trader is noise. A $500,000 purchase from a member who rarely trades is signal
Repeat accumulation Same member buying the same ticker across multiple disclosure periods Sustained accumulation over months expresses a thesis with duration -- exactly the kind of multi-month tailwind that longer-dated options strategies are built to capture
Sector context Policy environment, pending legislation, or contract cycle aligns with the sector A defence buying cluster during a geopolitical escalation period has a narrative. The same cluster without context requires more caution
Committee-Sector Alignment Reference

Armed Services / Intelligence

Defence contractors (Lockheed, Raytheon, Northrop, L3Harris), cybersecurity firms, satellite and communications companies. Members with classified briefing access on procurement cycles and threat assessments.

Finance / Banking

Major banks, regional banks, insurance companies, fintech. Members aware of regulatory direction, stress test outcomes, and interest rate policy discussions ahead of public announcement.

Energy / Environment

Oil and gas producers, pipeline operators, utilities, clean energy. Members aware of EPA rulemaking, pipeline approval timelines, and subsidy or tax credit legislation before public release.

Health / HELP Committee

Pharmaceutical companies, biotech, hospital systems, health insurers. Members aware of FDA regulatory posture, drug pricing legislation, and Medicare/Medicaid reimbursement changes.

Commerce / Technology

Semiconductors, major tech platforms, telecom. Members aware of antitrust direction, export control changes, and spectrum allocation decisions.

Strategy Selection by Signal Quality

The right strategy depends on how strong the signal is and how much time has passed since the disclosure. Use this framework to match conviction level to structure.

High Conviction -- Cluster Signal, Committee-Aligned, Large Position

LEAPS Calls (6-18 Month Expiry)

The strongest fit for congressional trading data. Congressional members hold positions for months to years -- they are not day traders. A LEAPS call (expiry 6 to 18 months out) gives the thesis time to develop without racing against time decay. Theta decay on LEAPS is slow in the early months, which means the disclosure lag barely matters -- you are not trying to catch the initial move, you are positioning alongside a sustained institutional accumulation. Strike selection: 5-10% out of the money gives meaningful leverage at a reasonable premium. Best applied when cluster buying is confirmed and the sector narrative is clear.

Bull Call Spreads (45-90 Day Expiry)

Buy a call at or near the money and sell a call at a higher strike with the same expiry. The sold call reduces your premium cost and your time decay exposure significantly compared to an outright long call. Maximum gain is capped at the spread width minus the premium paid, but maximum loss is also capped at the premium paid -- defined risk on both sides. A good default structure for medium-confidence signals where you want directional exposure but not the full premium cost of a LEAPS position. Use 45-90 day expiries to give the thesis room without excessive theta decay.

Diagonal Spreads (LEAPS Long, Near-Term Short)

Buy a LEAPS call as the core position and sell a shorter-dated call against it each month to reduce cost basis over time. Effectively a rolling income strategy on top of a long-dated directional bet. More active to manage than a straight LEAPS purchase but meaningfully more capital-efficient -- the monthly premium collected from the short call offsets the LEAPS cost over time. Best suited for high-conviction positions you intend to hold for multiple months while the congressional thesis plays out.

Medium Conviction -- Single Member, Reasonable Size, Sector Context Present

Short Puts (Cash-Secured, 30-60 Day Expiry)

Sell a put at a strike 5-15% below the current price, collect the premium, and repeat if the option expires worthless. If the stock drops to your strike and you are assigned shares, you own them at an effective cost below the current price -- and on a stock with confirmed congressional accumulation, that may be exactly where you want to be. Strike selection matters: go far enough out of the money that assignment is the exception, not the expectation. Requires sufficient cash in the account to cover the full share purchase if assigned.

Covered Calls on Assigned Shares

If you are assigned shares via a short put, hold them and sell calls against them rather than exiting immediately. The congressional accumulation data gives you a reason to maintain the position while generating income via the call premium. Strike selection is critical: go far enough out of the money to give the stock room to run if the thesis plays out. Aggressive covered calls (near-the-money strikes) cap your upside on a position where the underlying thesis may have significant remaining movement. Use 30-45 day expiries and strikes at least 10% above your cost basis.

Lower Conviction or Income Focus

Bull Put Spreads (30-45 Day Expiry)

Sell a put and buy a lower-strike put as protection, collecting a net premium. Defined risk on both sides -- you know your maximum loss before entering. Good for situations where the congressional signal is present but not strong enough to justify outright directional exposure. Generates income when the stock stays above your short strike, which is most of the time on a stock with informed accumulation behind it. Requires less capital than a cash-secured put because the long put acts as a margin offset.

Position Sizing and Portfolio Construction

Congressional signals are probabilistic. Options add leverage. These two facts together mean position sizing is not optional -- it is where the strategy succeeds or fails over time.

Practical sizing framework: No single position should exceed 10-15% of your total options allocation. This gives you room for 6-8 simultaneous positions without any one trade being catastrophic. Defined-risk structures (spreads, LEAPS calls) are more forgiving of sizing errors than short puts. LEAPS positions can reasonably carry a slightly larger allocation than short-dated plays because time decay pressure is lower and there is more time to manage the position if the thesis develops slowly.

Check sector concentration before adding any new position. Congressional cluster signals often cluster in specific sectors simultaneously -- defence during geopolitical tension, pharma around healthcare legislation. A portfolio built entirely on congressional signals can end up 60-70% concentrated in one sector without the trader noticing. Before entering a new position, look at your existing holdings and ask whether the new trade adds concentration risk you are not being paid to take.

Managing Positions -- What Kills This Strategy

Earnings and Binary Events

Implied volatility expands into earnings announcements and binary regulatory decisions (FDA approvals, contract awards, tariff announcements) and collapses immediately after -- regardless of which direction the stock moves. This IV crush destroys option value even when the underlying move is in your favour. Check earnings dates before entering any position and either close or roll before the event, or size the position knowing the IV collapse is coming. Congressional thesis trades are medium-to-long-term by nature -- binary events are noise unless the thesis is specifically tied to the outcome.

Low Liquidity on the Options Chain

Not every stock with congressional accumulation has a liquid options market. Wide bid-ask spreads on options chains are a hidden tax that compounds across multiple trades. Before entering a position, check the bid-ask spread on the specific contract -- anything wider than $0.20-0.30 on a contract priced under $5 should be treated with caution. Stick to higher-volume underlyings (large-cap names with active options markets) when possible, especially for spread strategies where you are trading two legs simultaneously.

Stale Signals and Already-Run Stocks

If a disclosure is 40 days old and the stock has already moved 25%, you are not getting in at the start of the move -- you are getting in after it. Evaluate the current price relative to both the disclosure date price and your thesis before entering. A stock that has already run hard may need time to consolidate before the next leg. LEAPS give you the most flexibility here because you are not racing the clock.

Sector Rotation

Congressional accumulation in a sector that subsequently rotates out of favour will underperform regardless of how sound the underlying thesis is. Market-wide sector rotation driven by interest rates, macro data, or sentiment shifts can overwhelm a well-reasoned individual position. Position size and defined-risk structures are the primary defences -- keep losses bounded and let the winners compound.

Claude as an Optional Tool

Claude can augment this workflow at several points without being required for any of them. Every step below can be executed manually using Capitol Trades, Quiver Quantitative, and your brokerage's options chain interface.

1

Screen for High-Quality Signals

Give Claude the Capitol Trades URL and ask it to identify recent cluster buys -- multiple members buying the same ticker within a 30-day window. Ask it to cross-reference committee assignments for each member and flag trades where the member's committee aligns with the sector of the underlying stock. This can be done manually but takes 20-30 minutes per scan. Claude reduces that to under a minute and can run it on a schedule if you connect it to the data source via MCP.

2

Evaluate the Options Chain

Once you have a candidate stock, ask Claude to evaluate the current options chain for that ticker -- implied volatility relative to historical volatility, available strikes and expiries that fit your strategy, the bid-ask spread on your target contracts, and whether any earnings or binary events fall within your intended holding period. Claude can do this if connected to a market data source, or you can paste the options chain data manually and ask for a structural recommendation.

3

Build the Trade Structure

Ask Claude to recommend a specific trade structure for a given signal -- which strategy type fits the conviction level, which strikes to use, which expiry, and what the maximum loss and maximum gain look like. Give it your account size and per-trade allocation cap so the recommendation is sized appropriately. Review its reasoning before executing. Claude should be treated as a structuring assistant here, not an autonomous decision-maker -- the final call is yours.

4

Track Positions and Manage Exits

Ask Claude to maintain a simple position log -- entry date, structure, strikes, expiry, premium paid or collected, thesis, and target exit. On a scheduled basis, ask it to flag any positions approaching expiry, any positions that have hit 50% of maximum profit (a common early-exit trigger to lock in gains and redeploy capital), and any positions where an earnings date is approaching within the holding window. This does not require automation -- a weekly manual check with Claude reviewing the log is enough for most position counts.

Past outperformance of congressional portfolios is documented in academic literature and in aggregate data published by services like Quiver Quantitative. It does not guarantee future results on any individual trade or position. The edge is probabilistic and requires correct application of the signal filters described above to be meaningful.
The STOCK Act's 45-day disclosure window means you are almost always trading on information that is at minimum several weeks old by the time it is actionable. LEAPS and longer-dated spreads are structurally better fits than short-dated options for this reason. Short-dated options (under 30 days) on congressional signals are generally too dependent on timing to work reliably after disclosure lag.
Some congressional members have faced scrutiny and investigation for trades that appeared to precede material non-public announcements. Following their disclosed trades as a retail investor using public data is legal. The disclosures themselves are public record. This is not insider trading -- it is reading public filings and acting on them, which is explicitly permitted.
Options approval levels vary by brokerage. Canadian residents using IBKR should confirm their options permission level covers the strategies they intend to use before building a position. Selling cash-secured puts requires Level 2. Spreads require Level 3. LEAPS purchases are covered under basic options approval.

Transcript

This transcript was auto-generated and may contain errors in speaker attribution, transcription accuracy, or formatting. Long transcripts may be truncated due to processing limits. Confirm accuracy and completeness against the original source before referencing or republishing.

[00:00]
Claude has changed how we trade stocks forever, and it's because of this new skill that lets it access live market data, track what Wall Street whales and US politicians are buying, and use that to play the stock market with all that information automatically. I've actually been using this for the past couple of weeks, and it's completely changed how I think about trading.
[00:30]
In this video I'm going to break down how you can get started in three levels. First, we're going to get set up -- I'll walk you through the basics of how trading works, and we'll get Claude and all the tech set up. Then we're going to build a copy trading bot -- tracking what Wall Street whales and US politicians are buying, and having Claude copy their exact moves automatically. In level three we're going to get into options -- what they are, how they work, and then we're going to build a bot that runs one of the most consistent income strategies in trading called the wheel strategy.
[00:58]
Before we get into anything, I need to give you some context on why I'm actually making this video and why this matters. For the first few years of my career I used to work at JP Morgan. And I actually got to see how institutional trading works up close. The biggest thing I took away from that is the gap between Wall Street and regular people comes down to just three things.
[01:20]
First, it's the data. Imagine a poker game, and you're sitting at a table with your two cards trying to figure out what to do. But the person across from you can see every card on the table and every card in your hand. You'd probably never sit down at that table, right? And that's the stock market. Wall Street knows when a billionaire places a massive bet on a stock. They know when a senator buys shares right before a major announcement. They see where the money is moving before you even hear about it on the news. And by the time you do hear about it, they've already made their money.
[02:00]
Second is execution. Say you spot the perfect trade, but you're at lunch or you hesitate for 10 minutes and the window closes. Wall Street doesn't deal with that. They have systems watching the market and placing trades around the clock automatically. And luckily, Claude can do that too.
[02:20]
Which brings us to number three, which is intelligence. Having the data and the speed really means nothing if you don't have a plan. The big guys have teams reading all that information and making calculated decisions. And to do all of that, you need to buy expensive tools. But with Claude and these new skills, that barrier is gone.
[02:41]
Alright, let's build this thing. If our goal is to get Claude to trade stocks for us, to do that Claude needs two things. It needs somewhere to place the trades, and it needs the information to decide what to trade. Back in the day, if you wanted to buy a stock, you'd have to call a person on the phone. That person would go to the stock exchange and make that happen for you. That whole system -- the firm, the person, the access to the exchange -- that's called a brokerage. Today it's an app. You open up Robinhood or Fidelity, tap a button, and you own a stock.
[03:10]
Some brokerages actually let you skip that step entirely and connect through code. And because Claude is so good at coding, that's what we need -- we're not going to be the ones trading, Claude is. We're going to be using a tool called Alpaca. It's free to sign up, and it gives you API access, which is how we're going to get Claude to connect to it and place trades on our behalf.
[03:46]
Quick disclaimer: I'm not a financial advisor. This is not financial advice. I'm just showing you how to build cool stuff with AI. And everything we're doing today is in paper trading -- using fake money in a real market. Same stocks, same prices, everything behaves the same way. You're just not risking a single dollar.
[04:10]
To get started, you need the Claude Desktop app -- claude.com/download. Download it for Windows or Mac. Make sure your OS is up to date. Ideally you have the Pro or Max version of Claude paid for.
[04:51]
Step two is the brokerage account. Search Alpaca, go to their site, and sign up. After you enter your information, you'll see your dashboard. We're going to be using the paper trading account -- not real money. You can create a new paper account, give it a starting balance (the video uses $50,000), and now you have capital to work with in a simulated environment.
[06:06]
To connect Alpaca to Claude, scroll down in your Alpaca dashboard to the API keys section, hit "Generate New API Keys," and you'll get three things: endpoint, key, and secret. We need all three.
[07:30]
Go back to the Claude app, switch to the Code interface, create a new project folder called "trading," and paste in your endpoint, key, and secret. Then tell Claude: "I just gave you the documentation and my keys to connect to my Alpaca trading account. Can you buy one share of Apple?" If the trade appears in your Alpaca dashboard, your connection is live. Tell Claude to save those credentials to a file in the trading folder so you don't have to re-enter them each session.
[09:27]
Right now Claude can trade, but it's trading blind. It doesn't know what to buy, when to buy, or why. It's missing the most important piece, which is strategy. In this level we're going to turn Claude into something that runs on its own -- a bot that watches the market, makes its own decisions, based on rules you set.
[10:05]
The first strategy is the trailing stop. Think of it like a smart thermostat. Once you set the rules, you don't have to sit there fiddling with it all day. The thermostat checks, acts, and adjusts. The bot works exactly the same way.
[10:15]
Here's how the trailing stop works. You buy a stock at $100. You tell Claude: if this drops to $95, sell it. That's your floor. $5 is the most you're willing to risk. Now if the stock climbs to $110, your floor is still $95 -- you're exposed to a $15 fall. That makes no sense. So what we want is Claude to move the floor up to $105. If the stock dips and hits that new floor, you're still up $5 in profit, and you were protected the whole time. The floor only goes up, never down.
[12:00]
The worst thing you can do is hand your AI a pile of money and say "go figure it out." The rules aren't the limitation. That's the whole point. This is how you take what you know -- your instincts, your risk tolerance, your read on the market -- and you encode it. Claude can execute thousands of decisions faster than any human. It's executing your decisions, just running at the speed and discipline you never could on your own.
[12:45]
To set this up, I'm in my Claude chat and I just speak to it. I want your help to schedule a trailing stop strategy on Tesla. Buy 10 shares at market price. Set a floor -- if it drops 10%, sell everything. If the stock goes up 10% from what I paid, move my stop loss up 5% below the current price. Every time it climbs, move the floor up again. And ladder in -- if it drops 20%, buy 10 more shares. If it drops 30%, buy 20 more. After you set this up, show me a summary of every order placed.
[15:01]
After the trades are placed, tell Claude to set a schedule during market hours to keep checking: "Hey, can you set up during market hours every day that you're checking consistently when we need to move our floor up or make new stop losses or re-enter. Use /schedule to make sure we have that going." Claude will create a cron job visible in your scheduled tasks panel. You can see it -- Tesla trailing stop monitor, running every 5 minutes Monday to Friday, 9:00 a.m. As long as your computer is on, it runs on its own.
[17:09]
Now that we have the trailing stop working, let's add ladder buys. Ask Claude to think through good ladder buy levels and update the strategy. It will produce something like: if the stock drops 15%, buy 10 shares; if it drops 30%, buy 50 shares. You can change these around, but the core idea is that as the price goes down, Claude is systematically buying in at better prices rather than just sitting on a loss.
[18:19]
We're still picking stocks ourselves. We chose Tesla because we like it -- that's a gut feeling. The biggest traders don't trade on gut feelings. They trade on information. In the next level we're going to show Claude where the smart money is.
[19:19]
On Wall Street, there are people who move millions of dollars in a single trade. When someone puts $50 million into a stock, they didn't do that off a gut feeling. They have research teams and private data. These people are called whales. And a group that does this a lot is, surprisingly, US politicians. Members of Congress are required by law to report their stock trades. The data shows that many of them consistently beat the market. They sit on committees, regulate entire industries, get briefed on policy changes before the public hears about them.
[20:00]
Think of MCP like a power outlet. The electricity -- the insider information -- is running through the walls, and we plug Claude into it. These services have data flowing through their systems, and MCP is the plug that connects Claude to their data. Once it's plugged in, Claude can pull from it anytime. Claude goes from being smart but working with nothing to being smart and seeing what the big players see.
[21:30]
For the copy trading setup: create a separate paper trading account in Alpaca, generate new keys, start a new Claude session, paste those credentials plus the Capitol Trades URL, and instruct Claude to find a politician with strong recent performance, monitor their disclosures, and copy those trades on a schedule. In the demo, Claude automatically selected Michael McCaul and explained its reasoning. The schedule was set immediately -- no manual configuration needed.
[25:14]
What is an option? An option is a contract between two people who agree to do a deal. You pay your car insurance company $100 a month. In exchange you get the right to file a claim if something goes wrong. If you crash, the insurance covers it. If nothing happens, they keep your $100. They're getting paid to take on the risk. Options work exactly the same way.
[26:23]
A call option gives you the right to buy a stock at a locked-in price. Say Apple is at $200, you buy a call with a strike of $210. If Apple goes to $230 before expiration, you can buy it at $210 and pocket the $20 difference. If Apple stays below $210, the contract expires and you lose the premium. Think of it like putting a $500 deposit on an apartment to lock in $2,000 rent for 30 days.
[27:06]
A put option gives you the right to sell a stock at a locked-in price. You own Apple at $200, you buy a put at a strike of $190. If Apple drops to $170, you can sell at $190 instead. The put protected you from the drop. That's the insurance side -- you paid a premium to protect yourself against a price decline.
[27:35]
When you sell an option, you become the insurance company. Someone pays you for a contract, and most of the time that contract expires without anything happening, and you keep the money. Insurance companies make billions doing this -- they collect premiums from millions of people and pay out on a small percentage of claims. The math works in their favor over time. Selling options works the same way.
[28:27]
The wheel strategy with Tesla at $250. You'd rather own it at $230. So you sell a put at the $230 strike -- for making that promise, someone hands you $5 per share ($500 total) right now. Two things can happen: Tesla stays above $230, the contract expires worthless, you keep the $500, and do it again next week. Or Tesla drops below $230, you have to buy it -- but you already collected $500, so your real cost is $225. You got Tesla cheaper than anyone else, and you wanted it anyway.
[29:30]
Stage two: you own 100 shares at an effective cost of $225. Now sell a covered call at the $260 strike. Someone pays you another $5 per share ($500) for the right to buy your Tesla at $260. If Tesla stays below $260, the contract expires, you keep the shares and the $500, and sell the call again next week. If Tesla goes above $260, your shares get sold: you bought at $225, sold at $260 ($35 gain per share), plus $5 from the put, plus $5 from the call -- $45 per share total, $4,500 on 100 shares. Then go back to stage one.
[31:32]
The wheel strategy sounds great until you try to run it yourself. You're picking strike prices based on market conditions every single week. Most people who learn about this give up after a few weeks because the management buries them. Claude handles all of it -- monitoring positions, picking expirations, rolling contracts when needed. You collect the premiums on a schedule and check in once a day.
[32:08]
The wheel strategy prompt given to Claude: run a wheel strategy on Tesla using my Alpaca account. Stage one: sell cash-secured puts with a strike around 10% below current price, two to four weeks out. If the put expires worthless, sell another. If assigned, move to stage two. Stage two: sell covered calls with a strike 10% above what I paid, same expiration window. Rules: never sell a put unless you have enough cash to buy the shares if assigned. Never sell a call below cost basis. Close any contract at 50% profit. Check positions every 15 minutes during market hours. Send a daily summary at market close.
[34:10]
If you can see the power of Claude but haven't used it before, or you've tried it but aren't sure how to get the most out of it, I'm putting together a full course walking through everything from scratch -- how to set up Claude, how to use it, how to build projects like this one. Stay subscribed for that. And watch the next video if you want to go deeper on any of these strategies or build something else together. Join the Claude Club where we can do this together and have some fun.

AI Master Prompt

The AI prompt on this page is auto-generated from the transcript content and is intended to support further exploration of the topics, concepts, and conclusions discussed. It is provided for informational purposes only. The user is solely responsible for all outcomes resulting from its use.

Master Prompt
You are an expert guide on using AI -- specifically Claude -- to automate rules-based stock market strategies using brokerage APIs, scheduled monitoring, and external data feeds. Your knowledge covers three interconnected domains: (1) automated trading bot construction using Claude Desktop connected to the Alpaca paper trading API, (2) copy trading using publicly disclosed congressional stock data via services like Capitol Trades and Claude's MCP integrations, and (3) options income strategies -- specifically the wheel strategy -- managed and executed by Claude on a scheduled basis. The framework is built around a single premise: Wall Street's edge over retail traders comes down to three things -- data access, execution speed, and analytical intelligence. Claude with the right integrations can close all three gaps simultaneously, at near-zero cost. The presenter, Samin Yasar (ex-JP Morgan, ex-Amazon engineer), demonstrates this across three levels of complexity: getting Claude to place basic trades via API, building a trailing stop and ladder-buy bot with cron-job scheduling, and running a full wheel strategy where Claude manages strike selection, expiration picking, and rolling decisions autonomously. Core principles of the approach: -- Rules encode the trader, not replace them. Claude executes your strategy -- your risk tolerance, your entry logic, your exit rules -- faster and more consistently than any human could. The rules are not the limitation; they are the point. -- Paper trading is the only sane starting point. All three strategies are demonstrated in Alpaca's simulated environment before any real capital is considered. -- Data advantage is the biggest edge. Congressional trade disclosures are public, machine-readable, and have historically outperformed the S&P 500. Plugging Claude into Capitol Trades via MCP is a practical, legal way to follow informed money. -- Selling options puts you in the house's position. Collecting premiums as the seller of puts and calls -- the wheel strategy -- generates income whether the stock goes up, down, or sideways. Management complexity is what kills most people who try this; Claude eliminates that problem. -- Cron-job scheduling makes the bot alive. Once rules are set and a schedule is created, Claude checks positions on its own during market hours. No manual intervention needed between sessions. What this is not: This is not a day-trading or prediction framework. None of the strategies involve trying to guess where a stock will go. The trailing stop manages downside automatically. The wheel collects income from volatility. The copy trading follows disclosed trades from people with structural information advantages. This is also not financial advice -- the presenter is explicit about that throughout. The framework is about encoding rules-based logic into Claude and letting it execute consistently. How to use this chat: 1. SETUP HELP: Ask for step-by-step guidance on connecting Claude Desktop to the Alpaca API, generating API keys, saving credentials, and confirming a first trade. Walk me through what I need and where I might get stuck. 2. STRATEGY DESIGN: Help me define or refine a set of trading rules -- trailing stop levels, ladder buy triggers, options strike selection logic, or wheel strategy parameters -- in plain English that I can give to Claude as a prompt. 3. PROMPT BUILDING: Help me construct the exact Claude prompt I would use to set up any of the three bots: trailing stop bot, copy trading bot, or wheel strategy bot. Make the prompt specific, complete, and immediately usable. 4. SCENARIO TESTING: Walk me through what the bot would do in a specific market scenario -- a sharp drop, a gap up, an assigned option, a stock halt -- so I can validate my rules before relying on them. 5. CONCEPTS: Explain any underlying concept in plain language -- trailing stops, ladder buys, call options, put options, covered calls, cash-secured puts, the wheel cycle, MCP, cron scheduling, paper trading -- at whatever depth I need. 6. RISK AND LIMITS: Help me think through the edge cases, failure modes, and practical limits of these approaches. What happens when the market is closed? What if a stock halts? What does "assigned" actually mean in practice? Respond in a grounded, direct, technically precise way -- no hype, no cheerleading, help me understand and build correctly. To start: What is your current setup -- do you have Claude Desktop installed, an Alpaca account created, or any part of this already running? Or are you starting from scratch?