Skip to content

Guide · Agent skill

How to use the Trade Journal Analysis agent skill (trade-journal)

Updated · written with AI from the project's own documentation and checked against our catalogue data

The trade-journal skill lets an AI agent read your broker's trade export, a CSV or Excel file, and describe your past trading: win rate, holding periods, drawdown, most-traded symbols and when you trade. It also checks four common behaviour patterns and gives each a severity level with numbers behind it. This guide covers what the skill does, how to install it, how to ask for an analysis and what to watch out for.

Trade Journal AnalysisHKUDS/Vibe-Trading · MIT

Quick start

  1. 1Download the ZIP. Get the trade-journal ZIP from this page.
  2. 2Unzip into a skills folder. Place it in ~/.claude/skills/ for personal use or .claude/skills/ inside your project.
  3. 3Or upload it in the apps. In the Claude apps, go to Settings → Capabilities → Skills and upload the ZIP.
  4. 4Confirm the tool is available. Make sure your Vibe-Trading agent environment has the analyze_trade_journal tool.
  5. 5Upload your broker export. Provide a CSV or Excel file exported from your broker.
  6. 6Ask for the analysis. Request a full review or a narrower view such as one month, one stock or one market.
  7. 7Review the diagnostics. Check the severity ratings and the numeric evidence before drawing conclusions.

What the skill makes the agent do

Agent skills are instruction packages that give an AI agent a specific capability. This one tells the agent how to analyse a trade journal using the analyze_trade_journal tool. Once you upload a broker export, the agent parses it and writes a single report in your language.

The skill covers two layers. The profile layer reports holding days, trade frequency, win rate, profit/loss ratio, cumulative profit and loss, maximum drawdown, top symbols and the market and hourly distribution of your trades. The behaviour layer scores four biases (disposition effect, overtrading, chasing momentum and anchoring), each rated low, medium or high with numeric evidence.

  • Supported formats are detected automatically: Tonghuashun (同花顺) and Eastmoney (东方财富) A-share CSVs, usually GBK-encoded, Futu (富途) HK/US CSVs in UTF-8, and a generic CSV with columns like datetime, symbol, side, qty and price.
  • Only closed round-trips count toward win rate and profit/loss ratio. Open positions are excluded.
  • Mixed-currency journals are reported per currency. The skill does not convert exchange rates.

Install the trade-journal skill

Download the ZIP for this skill from this page. Unzip it into one of these folders: ~/.claude/skills/ for personal use across projects, or .claude/skills/ inside a specific project folder if you want it limited to that project.

If you use the Claude apps instead, open Settings, go to Capabilities, then Skills, and upload the ZIP there.

  • Personal install: unzip into ~/.claude/skills/
  • Project install: unzip into .claude/skills/ in your project
  • Claude apps: Settings → Capabilities → Skills, then upload the ZIP

Check the requirements before you start

The skill needs the Vibe-Trading agent environment, because it calls the analyze_trade_journal tool that lives there. Without that tool available, the skill cannot parse anything. You also need a CSV or Excel broker export that you upload yourself.

  • The Vibe-Trading agent environment with the analyze_trade_journal tool available
  • A CSV or Excel broker export file uploaded by you
  • The skill runs scripts; it does not use the network

Ask for an analysis in plain language

You do not need special syntax. Upload your file and ask in normal words, for example for a full review or only one part. Behind the scenes the agent calls analyze_trade_journal with the file path, and can narrow the analysis with a date range, a single symbol or a market.

Useful requests include a general review, a profile only, behaviour diagnostics only, a time slice, a single stock, or one market. The agent should not re-upload the file for follow-up questions in the same session.

  • Full review: "Analyse my trade journal" (runs profile and behaviour)
  • Profile only: a fast summary of metrics
  • Time slice: "How did I do in March?" (date range filter)
  • Single stock: "How much did I make on Moutai?" (symbol filter, e.g. 600519.SH)
  • Market split: "Show Hong Kong and US stocks separately" (market filter: hk, us, china_a or crypto)
analyze_trade_journal(file_path="uploads/xxx.csv") analyze_trade_journal(file_path="uploads/xxx.csv", analysis_type="profile") analyze_trade_journal(file_path="uploads/xxx.csv", filter_expr="2026-01 to 2026-03") analyze_trade_journal(file_path="uploads/xxx.csv", filter_expr="symbol=600519.SH") analyze_trade_journal(file_path="uploads/xxx.csv", filter_expr="market=china_a")

Understand the four behaviour diagnostics

Each diagnostic compares a measurable pattern in your history against a threshold. The thresholds are heuristics, so a high rating means the pattern is strong in your data, not that you have a proven bias.

Disposition effect compares how long you hold losers against winners. Overtrading compares your average daily profit on busy days (3 or more trades) with quiet days (1 or fewer). Chasing momentum measures how often you buy after a rapid run-up of more than 3% in the same symbol. Anchoring checks whether frequently traded symbols stayed in a narrow price band.

  • Disposition effect: loser hold time divided by winner hold time; medium at 1.2 or more, high at 1.5 or more
  • Overtrading: gap between quiet and busy day P&L; medium at 0.3 or more, high at 1.0 or more
  • Chasing momentum: share of buys after a 3% run-up; medium at 40% or more, high at 60% or more
  • Anchoring: share of symbols traded five or more times with price variation under 5%; medium at 33% or more, high at 66% or more

Read the report and act on it

The report leads with top-line numbers, then the top five symbols, market split, trading hours and a one-sentence observation. If your win rate is below 40% and your profit/loss ratio is below 1.0, the agent will say so explicitly. Very short holding periods combined with many trades per week are flagged as an intraday pattern.

For any diagnostic rated medium or high, the report suggests a practical rule to test, such as a fixed stop-loss, a daily trade limit, or buying pullbacks instead of new highs. These are suggestions to consider, not instructions.

What people use it for

  • Reviewing a year of A-share trades exported from Tonghuashun to see whether you hold losers longer than winners.
  • Comparing March performance against January and February using a date-range filter.
  • Checking how much you earned or lost on one stock, such as a single symbol, across all its round-trips.
  • Studying trading behaviour from broker exports as part of a course or research project.
  • Splitting a Futu export into Hong Kong and US market views to compare them separately.

Frequently asked questions

How do I install the trade-journal agent skill?

Download the ZIP from this page and unzip it into ~/.claude/skills/ for personal use, or into .claude/skills/ in a project folder. In the Claude apps, you can instead upload the ZIP under Settings → Capabilities → Skills.

What do I need to run the trade-journal skill?

You need the Vibe-Trading agent environment with the analyze_trade_journal tool available, plus a CSV or Excel broker export that you upload. The skill runs scripts but does not use the network.

Which broker formats does it support?

It auto-detects Tonghuashun (同花顺) and Eastmoney (东方财富) A-share CSVs, usually GBK-encoded, Futu (富途) HK/US CSVs in UTF-8, and a generic CSV with datetime, symbol, side, quantity and price columns.

Is the trade-journal skill free?

The source material for this page does not state pricing or licence terms beyond the listed MIT licence, so check the repository's licence file for the full terms.

Why does the skill say I have a bias when I think I do not?

The four diagnostics are statistical heuristics that compare patterns in your records against fixed thresholds. A high rating shows the pattern is strong in your data, not that it is a proven psychological bias, so treat it as a prompt for review.

Can it convert currencies or backtest a strategy?

No. For mixed-currency accounts it reports each currency separately without FX conversion, and strategy extraction and backtesting are planned for a later phase and are not available yet.

New agent skills, repos and AI tools — every week

The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.

Our verdict

Use this skill if you want an honest, data-based review of your own past trades from a Chinese or international broker export, or you are studying trading behaviour. Pick something else if you want investment advice, return predictions, combined multi-currency totals, or backtesting, since the skill does none of those yet.

Download the skill