Please explain me how your (Claude) billing system works: I see this in https://claude.ai/settings/billing: $252.87 Current balance and here https://platform.claude.com/settings/billing $11.08Remaining balance I use Claude Code Terminal, Claude API with a Key and Claude Chat window in here. Please write me a complete prompt (Claude Code) to analyze current code and fix the problem to show these values on market screen in this order and Realtime values on the time of the request. S&P 500 (^GSPC) Dow Jones Industrial Average (^DJI) NASDAQ Composite (^IXIC) Russell 2000 Index (^RUT) CBOE Volatility Index (^VIX) Bitcoin USD Price (BTC-USD) Oil Fed Red EUR/USD These dates must also be retrieved in RealTime (not from Cache) for the Market Evalution. Please analyze code throughly and fix any problems. Ultrathink Please write me a detailed Prompt for Claude Code Terminal as a Agent thats achieves the follwing: Make design changes in the UI of the Flutter App. Goal: Make it more userfriendly and better looking Where: Login Screen 1) Add all interesting dates like VWAP and other on the screen 2) After analyzing and EXIT is recommended a Cnonfirmation prompt is shown. Please just show the recommendation and remove the prompt. Don't forget to update technical documents like Overview, Build Summary and Tasks (if some pending) Where: Markets Screen Where: Favourites Sreen Where: Deep Dive Result Screen Where: Stock Radar Candidates Screen Please read files Docs/TradeBot. Check Phase 10 of the TradeBotTodoList with the build_Summary and implement all pending task of the phase 7. Update the todo list and Buid Summary. Please write me a detailed Prompt for Claude Code Terminal as a Agent thats achieves the phase 11 of the TradeBotTodolist and update technical documents like Overview, Build Summary and Tasks list and At the end a Summary. Please read files Docs/TradeBot/*.* Please give me an overview how the tradebot will work. 1) Which are the schedules of each Tasks 2) How the bot will get the stocks that are possible candidates for trading 3) What strategies will the bot use Please read @Docs Please execute tasks of phase 10 (update Build Summary + Todolist). Give an summary at the end. Please read all file in @Docs\TradeBot and create a file with a Standard Operating Procedure (SOP). How they work together The skills cross-reference each other intentionally: new-flutter-feature ends by pointing at new-riverpod-provider and update-aiticker-docs. new-riverpod-provider and new-rest-endpoint both end by pointing at update-aiticker-docs. write-claude-code-prompt references update-aiticker-docs so every Claude Code prompt always includes the doc-update step in section 3. How to use them The agent will pick a skill automatically whenever your request matches a skill's description. A few example phrasings that will now route to a skill: "Write a Claude Code prompt to add a notifications tab" → write-claude-code-prompt "Scaffold an alerts feature in the Flutter app" → new-flutter-feature (+ new-riverpo d-provider) "Add a POST /api/alerts/create endpoint" → new-rest-endpoint "I just changed code in rest-api/api/" → update-aiticker-docs Please prepare me a basic Flow Diagram file of the tradebot DOcs/Tradebot/flow.md . News fetched from Sources defined in CI Sources with frecuency indicated Entities added to News article (frecuency?) Candidates fetched from News Signals? Scores? Tradebot activated? Please improve this screen "CI -> News" for light and dark Please write a complete and detailed prompt (Handoff) for Claude Code to implement these changes and improvements. Please read all file in @Docs Please check and compare the Overview and current Implementation with my desired functioning: - We have news source that produce news articles - The news articles are evaluated and we get Entities and signals We create or have TradeBots A TradeBot has the following parameter: if activated (we should also add a timeframe when he is running) Scan interval, Monitor Interval, Max Open Positions, Max trades per Day, Daily loss limit, same-symbol cooldown Position sizing, Position Size valuem Entry slippage, Exit slippage, allowed side, market Scope Allow Overnight, EOD flatten, Paper Mode Monitor mode, Execution backend, Stop-loss-mode, Trail amount, Syop-loss routing, Strategy profile Linked Alpaca Account A active Tradebot should loop thru the entities and signals and Deep Dive each Symbol until he finds a candidate for trading: Start trading->watch trade -> exit Trade -> Report Result -> Get next symbol Let me know if you understand my idea of the tradebot. Please explain differences (simple) and reasons. Than we can decide changes and improvements. Please improve this Trade Detail Overview Screen of the AITicker Admin APP for Dark and Light and write Claude Code Terminal Prompt (Handoff) in a downloadable md file 1) Please fix the broker submission 2) Where can I set owner_max_trades_per_day? 3) Please check: Sanity-check why paper_trades 7 & 8 fill locally with no alpaca_order_id, while 9–11 hit Alpaca — there's an inconsistency in the execution path. Please add a new menu item Settings to the TradeBot menu with a settings screen where I can edit the trade_bot_owner_settings. Please add to these settings a section to add another Claude API keys (I want them separated from the AI Ticker) I would like to have a new Trades Section in the menu between Dashboard and Settings with a List View of all Trades (with Filter) and a detail view. The list view has the same columns as the "Recent Paper Trades (all bots)". The detail view should show all data related to a trade, entry price, stop/losses set, exit price, all decisions (prompts), any other data related. Also with a button "Analyze Trade for Learning" if not already run and if the trade is closed. Please ask questions if something is not clear. I see we have a lot of tests like the one shown above (rest-api\tests\phase10_7\PromptRenderServiceTest.php). I want to create a Test Page so I can run a tests and see the results. Please create the page with all existing tests. Some minor things to fix: News Page: Please fix the date (or since) as it always says 2 hours as the newest article (check time setting) TradeBot: We are currently in the testing phase. Yesterday we only had 5 trades. I want more at the moment to check if everything is running well. When all is running smoothly we can adjust prompts and other settings but now I need more trades. How to adjust to get more trades? Recommended next steps 1. Verify Phase 17.5 close-path fix is actually deployed (closeTradeViaAlpacaPaper, AlpacaOrderIngester). Today shows the exact stuck-open / 404-flood pattern that fix targeted. 2. Add a minimum-hold floor to the AI monitor (e.g. ignore exit recommendations within first N seconds / before X% adverse move). v3 is suiciding trades 12–36 s after fill. I Fixed scan overlap: bump scan_interval_minutes to 3. Please Investigate the 305 enter → 14 traded funnel — large drop-off should be intentional and visible. Please add a Decision Section where I can Monitor all decisions made with the cost. Also add a cost column to the decision list and to the list view of the trades. Please add 2 more tiles to the dashboard with Todays AI Cost and Montly AI Costs Please check Alpaca data: Some stocks have more than 1.5% gain but the Bot is still holding. How much are 7.5 bps? Please check the daily AI Cost, The value was higher previously in the day and than dropped to a lower amount. Also the usage in Claude API shows more usage. If the problem was caused by an update ignore it. Only make sure that the value is correct in the future. I want to automate the current process that I do now to analyze and compare data of the tradebot and alpaca to detect problems and errors: - Downloading Trade Log Export in the app - Posting into chat window - Log into Aplpaca Account and make/dawnloading Screenshot of todays (or last) activities - Post image in chat window. - Ask you to compare and analyze both data Questions: Can you create a script on the server so you can retrieve the Trade Log export directly? Does the alpaca api offer the activities and can you also create a script on the server to retrieve directly? If both is possible can I just ask you: Please check trade logs (or someting else) check trade logs I would like to work on the training/learning process: Some questions please answer and make suggestions: After what time cycle (day, week, month) we can use the data for a new strategy? As we currently have only one strategy I would think to improve the current with the post-mortem analysis and create a new one or? Any suggestions how we can do the process? What should be tackle next of both TodoLists in Tradebot. Please also create a Cost Screen (Menu item after Settings) where I can see where AI costs are generated. A list with the details of the usage if it came from the AI call of: News evaluation, Signals, Scores, entry, monitor or exit (any other also) TradeBot: Regarding AI Model change for entry/monitor/exit. I see that costs gone up when used Sonnet 4.7 for entry decsision. I changed to OPus 4.6 for now. I have another idea: First a prompt for Haiku and this model decides if we need to escalate to Sonnet 4.7 model? Is this possible? Tradebot: Please check if we have short sales enabled. I see the current bot is set to Allowed_side=long_short but I don't see short decisions. If disabled hardcoded please activate and let me know if we need anything else? 1) Important: Do we track the side short/long on every trade? If yes please add this column to the trades screens (Dashboard, Trades) 2) Please also record the Triage calls for AI cost tracking (I don't see them in the cost screen) Today, the TradeBot performed very bad, and he had almost five percent loss on the investment, and we need to check how the process works. First, we need to also learn from the AITicker Co-Pilot, so we need to add some basic parameters before we check any Ticker (saves AI Calls for entry, = costs). Also, the prompts need to be improved. We need to analyze which is the most important prompt (Entry, Monitor or Exit). The goal is to take small profits, like one percent on each trade. That's the goal. I don't know if the prompts are set to reach this goal. The AI should use known patterns in the candle sticks (like Hammer, reversed Hammer etc) and recommend entry, I don't see that the monitor knows what pattern did the entry response saw. The entry response should pass more information to the monitor prompt like we saw a Hammer Pattern so price should go up to target x. The entry prompt should gives us a probility if the stoch goes 1% up or down (thats our goal). Also we need to control costs, but on the other side, we need a better performance. Please help me to analyze and make suggestions what we can change. Please make sure the the trade monitor first ckecks the current price and compares to the target price (set by the AI or entry x goal +1%?) No AI Call required. If not make the ai call for hold or exit with loss. Do we use the origin market news and fundamentals (in the prompt) when the AI evaluates a stock for entry (We do this in the AiTicker CoPilot)? Please confirm that we now pass the entry evaluation to the next prompt (Monitor and Exit). Start Phase 7 from Docs/TradeBotImprovementTodoList.md. Same conventions as Phase 1+2+3+4+5+6 — see project memories and the Phase 1+2+3+4+5+6 entry in Build_Summary.md. Please write me an brief (but understandable) description of the current DayTrading process Start with having a Candidate (Ticker Symbol), evaluating it, how to enter a trade, how is monitoring work, when the trade will be exited for a external validation and more suggestions how to improve.