| name | trade-journal-shadow-review |
| description | Use when users need broker trade journal parsing, trading behavior diagnostics, FIFO roundtrip analysis, Shadow Account profile extraction, or post-trade review for 同花顺, 东方财富, 富途, generic CSV, or Excel exports. This skill uses trading_core journal adapters and Vibe-Trading-inspired Shadow Account concepts to produce win rate, profit/loss ratio, holding days, drawdown, overtrading, disposition effect, chasing momentum, anchoring, local .research artifacts, and Not investment advice without predicting returns or creating live trades. |
Trade Journal Shadow Review
Use this skill when the user asks to review personal trading records, parse a
broker export, diagnose behavior patterns, or create a Shadow Account profile
for simulation.
Required Order
financial-services-skill-router
.agents/SKILLS_INDEX.md
china-market-overlay when the records include A-share, 港股, iFinD,
同花顺, or AKShare context
- This skill
financial-output-qa-gate before delivery or persistence
Workflow Recipes And References
- Use
.agents/workflows/trade_journal_shadow_review.json for productized
broker journal review.
- Use
.agents/references/output-contract.md for shared output fields.
- Use
.agents/references/display-profiles.md for App/CLI/JSON presentation.
- Use
.agents/references/local-research-artifacts.md for .research/
artifact paths.
Core Command
python3 -m trading_core.cli journal --file uploads/trades.xlsx
Supported inputs: .csv, .xlsx, .xls.
CLI output defaults to a Markdown review card for readability. Add
--format json when the full normalized object is needed for automation.
App Conversation Display
When presenting results directly in the Codex app conversation, read
references/app-conversation-display.md and use the card-style Markdown layout:
one-line orientation, review card, behavior diagnostics table, Shadow Account
table, next review checklist, and compact data/QA section. Keep parser
assumptions, evidence gaps, and Not investment advice; do not show raw JSON
unless the user asks for it.
Output Objects
profile: trade count, closed roundtrips, average holding days, trade
frequency, win rate, profit/loss ratio, total PnL, and top symbols.
behavior_diagnostics: disposition effect, overtrading, chasing momentum,
anchoring, and evidence or source gaps.
shadow_account_profile: extracted rules from profitable closed roundtrips
for simulation and review only.
- Local artifacts:
.research/journals/*.analysis.json
.research/shadow/*.shadow.json
- QA gate:
financial-output-qa-gate
python3 tools/check_research_integrity.py --input path/to/journal.json --profile journal_shadow
Guardrails
- Do not infer missing trades, fees, prices, or quantities.
- Do not promise future performance from historical behavior.
- Do not generate personal position sizing or live order instructions.
- Label parsing limitations and any missing broker columns.