agentleFS
tusharagg1tusharagg1/aifolimizer/.claude/skills/trade-journal/SKILL.md

trade-journal

Skill2 starsChanged 5 months agoMITView on GitHub

Capture the psychological side of a trade - emotion, conviction source, confidence, and plan adherence - at entry and exit, then surface which felt-states actually lose money. Use when the user says "journal this trade", "log my entry", "I just bought/sold X", "reconcile my trade", "trade journal", "why do I keep losing", "what's my real psychological pattern", or wants emotional/behavioral review of their trading. Pairs with pre-trade-check (entry) and weekly-mirror (review).

10 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 3 are distinctive, and 2 are very specific to this project.

  • 0

    Common

    Most of the 9 edited versions we read keep these.

  • 3

    Distinctive

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Step 3 - Mirror back one line: If emotion ∈ {fomo, revenge, fear} OR conviction_source ∈ {tip, social, gut} OR …

    line 46
  2. Very specific:

    If it returns reconciled=false, there was no open entry - tell the user the entry was never journaled.

    line 66
  3. Distinctive:

    Never invent felt-state. If the user did not state an emotion, ask.

    line 104
  4. Distinctive:

    Keep entry/exit interactions to <120 words.

    line 106
  5. Distinctive:

    No PII. Journal stores ticker + felt-state only.

    line 107

“Common” is measured: kept by at least 6 in 10 of the 9 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Trade Journal (Psychological Layer)

Goal

Most retail losses are emotional, not analytical. decision_memory records the facts of a trade (price, conviction label, thesis) and shadow_account derives biases from price/date. Neither captures what the trader felt at the moment of entry. This skill does - emotion, conviction source, self-rated confidence, plan adherence - then get_journal_insights cross-tabs felt-state against realized outcomes to answer the one question that beats technical analysis: what emotional state precedes my losing trades?

This is a capture + mirror skill, not a recommender. It does not approve or reject trades (that's pre-trade-check).

When to invoke

  • ENTRY: user just placed (or is about to place) a trade and wants it journaled. Natural follow-on from pre-trade-check.
  • EXIT: user closed a position - reconcile the open journal entry.
  • REVIEW: "why do I keep losing", "what's my pattern", "trade journal insights".

How to run

Step 0 - Detect phase (REQUIRED): entry, exit, or review. If ambiguous, ask.

ENTRY phase

Step 1 - Capture felt-state (ask only what the user has not already said):

  1. Ticker?
  2. Emotion right now? → one of: calm | fomo | fear | revenge | conviction | bored | uncertain
  3. Where did the conviction come from? → thesis | chart | tip | social | news | gut
  4. Confidence 1-5? (1 = a flyer, 5 = highest conviction)
  5. What is the plan? (entry / stop / exit stated BEFORE the trade - one sentence)
  6. Did pre-trade-check pass? (yes/no - if they skipped it, note that, it is itself a signal)

Step 2 - Persist:

Call mcp__aifolimizer__log_trade_journal with:
  ticker=<TICKER>
  emotion=<emotion>
  conviction_source=<source>
  confidence_1to5=<1-5>
  plan_intended=<one-sentence plan>
  felt_note=<free text, optional>
  pre_trade_check_passed=<true/false>

Step 3 - Mirror back one line. If emotion ∈ {fomo, revenge, fear} OR conviction_source ∈ {tip, social, gut} OR pre_trade_check_passed=false, flag it plainly: "Logged. Note: this is a {emotion}/{source} entry - historically your weakest setup (check insights)." Do not moralize beyond one line.

EXIT phase

Step 1 - Reconcile (ask):

  1. Ticker?
  2. Did you follow the plan you wrote at entry? (yes/no - be honest)
  3. Emotion at exit? (same enum)
  4. Was the outcome a surprise? → expected | mild_surprise | shock
  5. One-line lesson?

Step 2 - Persist:

Call mcp__aifolimizer__log_trade_journal_exit with:
  ticker=<TICKER>
  plan_followed=<true/false>
  exit_emotion=<emotion>
  outcome_surprise=<level>
  lesson=<one line>

If it returns reconciled=false, there was no open entry - tell the user the entry was never journaled; offer to log a retroactive entry.

REVIEW phase

Step 1 - Mark outcomes current, then pull insights (in order):

  1. mcp__aifolimizer__resolve_trade_outcomes (marks decisions to market so wins/losses are fresh)
  2. mcp__aifolimizer__get_journal_insights

Step 2 - Present the mirror (use the output template). Lead with the single worst felt-state by win-rate. State the confidence-calibration verdict bluntly: if avg_confidence_losses >= avg_confidence_wins, the user's confidence is inverted - they feel most sure exactly when they are most wrong.

Output template (REVIEW)

TRADE JOURNAL - PSYCHOLOGICAL MIRROR
Scored trades: <scored_entries> of <total_entries>  (<open_unreconciled> not yet closed)

WIN-RATE BY ENTRY EMOTION
  conviction   <n>  <win_rate_pct>%
  calm         <n>  <win_rate_pct>%
  fomo         <n>  <win_rate_pct>%   <-- worst
  ...

WIN-RATE BY CONVICTION SOURCE
  thesis       <n>  <win_rate_pct>%
  social       <n>  <win_rate_pct>%   <-- worst
  ...

CONFIDENCE CALIBRATION
  avg confidence on WINS:   <avg_confidence_wins>
  avg confidence on LOSSES: <avg_confidence_losses>
  verdict: <calibrated | INVERTED - you are most sure when most wrong>

THE PATTERN: <one sentence naming the emotion+source that loses money>

Rules

  • Capture only - never approve/reject a trade (route to pre-trade-check for that).
  • Never invent felt-state. If the user did not state an emotion, ask; do not guess it from price action.
  • Enums are fixed - map the user's words to the nearest enum value; if none fits, ask.
  • Keep entry/exit interactions to <120 words. The review is the only long output.
  • No PII. Journal stores ticker + felt-state only; no balances, no account IDs.

Gotchas

  • get_journal_insights can only score entries whose ticker has a resolved decision (target_hit/stop_hit). If scored_entries is 0 but total_entries > 0, tell the user outcomes aren't resolved yet - run resolve_trade_outcomes or wait for trades to hit target/stop.
  • Insights need a decision logged too (via pre-trade-check / log_trade_decision). A journal entry with no matching decision is captured but unscoreable. Encourage logging both.
  • Small-sample caveat: with <5 scored trades per bucket, win-rates are noise. Say so.
  • This skill feeds weekly-mirror (lessons) - do not duplicate the performance-math output; defer P&L/R-multiple to weekly-mirror.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.