Plan & Review turns timestamped market snapshots into validated pre-market plans, intraday reviews, post-market analysis, and trading discipline scores.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install plan-review
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install plan-review using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Plan & Review?
Plan & Review is a read-only market research and trading review workflow for Openclaw Skills. It organizes the trading day into three stages: pre-market preparation, midday or intraday review, and post-market review. Each report uses validated market data, explicit hypotheses, evidence citations, counter-evidence, and structured conclusions rather than unsupported predictions.
The skill also maintains an auditable trading journal and supports disciplined topic lifecycle analysis, including observation, launch, expansion, climax, divergence, and retreat. Python workflow commands capture source snapshots, prepare phase-specific data bundles, generate derived metrics, and validate reports before delivery.
Plan & Review Use Cases
- Build an evidence-based plan before the market opens using the prior session, overnight context, market breadth, index history, sector flows, and watchlist data.
- Review midday market behavior and determine whether pre-market hypotheses are confirmed, partially confirmed, invalidated, or unsupported.
- Analyze the full trading session after the close and create a next-day research watchlist.
- Track theme and sector lifecycle changes across breadth, strength, persistence, fund-flow consistency, and counter-evidence.
- Maintain timestamped trading journals with evidence IDs such as
[E01]for every quantitative statement. - Apply strict data-quality gates when snapshots are missing, stale, inconsistent, incomplete, or affected by schema changes.
- Use Openclaw Skills for reproducible market research without delegating trade execution or relying on execution-layer tools.
How Plan & Review Works
- Capture timestamped source snapshots during the permitted Asia/Shanghai collection windows: midday from 11:30 to 13:00, market close after 15:05, and Dragon-Tiger List data after 16:30.
- Run
workflow.py preparefor the requested phase. The workflow checks required inputs, dates, coverage, and data freshness before returningstatus: ready. - Read the generated phase data bundle and the corresponding report template. The agent must use the bundle's actual market breadth, rankings, index history, limit-up and limit-down activity, ladder structure, sector distribution, and 1/3/5-day fund-flow data.
- For the pre-market phase, create a verifiable observation framework with no more than three daily hypotheses and explicit invalidation conditions.
- For the intraday phase, compare the midday snapshot with the previous trading day's close and evaluate each pre-market hypothesis using the required evidence categories.
- For the post-market phase, compare midday and closing snapshots, incorporate Dragon-Tiger List data, assess structural evolution, and create a next-day observation list.
- Classify themes using six lifecycle stages: observation, launch, expansion, climax, divergence, or retreat. Report breadth, strength, persistence, fund-flow consistency, and counter-evidence for each classification.
- Write only the template-defined sections and fields. Separate facts, interpretations, hypotheses, and counter-evidence; never infer historical values from a later snapshot.
- Run
workflow.py validatefor the selected phase. Do not deliver a report if validation fails.
Plan & Review Setup
Run all commands from the project root. The workflow scripts use Python's standard library, while market data collection uses the market skill environment.
uv venv market/.venv
uv pip install --python market/.venv -r market/requirements.txt
Capture the required market snapshots for a trading date:
plan-review/.venv/bin/python plan-review/scripts/workflow.py capture --phase noon --date YYYYMMDD
plan-review/.venv/bin/python plan-review/scripts/workflow.py capture --phase close --date YYYYMMDD
plan-review/.venv/bin/python plan-review/scripts/workflow.py capture --phase lhb --date YYYYMMDD
If plan-review/.venv is not available, run workflow.py with system Python; it launches subprocesses that use market/.venv for collection. Prepare each report only after its required snapshots are available:
plan-review/.venv/bin/python plan-review/scripts/workflow.py prepare --phase pre --date YYYYMMDD
plan-review/.venv/bin/python plan-review/scripts/workflow.py prepare --phase noon --date YYYYMMDD
plan-review/.venv/bin/python plan-review/scripts/workflow.py prepare --phase post --date YYYYMMDD
Create reports from the returned data bundle and templates, then validate them:
plan-review/.venv/bin/python plan-review/scripts/workflow.py validate --phase pre --date YYYYMMDD
plan-review/.venv/bin/python plan-review/scripts/workflow.py validate --phase noon --date YYYYMMDD
plan-review/.venv/bin/python plan-review/scripts/workflow.py validate --phase post --date YYYYMMDD
Optional watchlist data is read from data/watchlist.json. An unconfigured watchlist is ignored; partially available symbols produce warnings, while complete watchlist data failure blocks the pre-market report.
Plan & Review Data Schema & Taxonomy
Storage layout
| Data type | Location | Purpose |
|---|---|---|
| Raw market snapshots | market/data/report_snapshots/YYYYMMDD/ |
Timestamped source evidence collected during market windows |
| Phase data bundles | plan-review/data/YYYYMMDD/ |
Validated inputs prepared for pre, noon, and post workflows |
| Daily reports | report/daily/YYYYMMDD_盘前计划.md, YYYYMMDD_盘中复盘.md, YYYYMMDD_盘后复盘.md |
Human-readable plans and review journals |
| Optional watchlist | data/watchlist.json |
User-configured stocks for additional historical and indicator checks |
Required evidence domains
- Full-market advance/decline breadth, rankings, turnover, and activity distribution.
- Multi-period index history and previous-session closing baselines.
- Limit-up, limit-down, and failed-breakout activity from the prior five trading days.
- Sector and concept distributions with 1/3/5-day fund-flow rankings.
- Index, A50, USD/CNY, completed U.S. market index data, and commodity market conditions where available.
- Dragon-Tiger List details and institutional statistics for post-market review.
- Optional watchlist K-lines and indicators when a watchlist is configured.
Metadata and validation taxonomy
- Each bundle must retain its
bundle_id, phase, date, source status, timestamps, and evidence numbering. - Quantitative claims must include an inline evidence reference such as
[E01]on the same line. - Derived metrics include the versioned Fear & Greed index
fg-v1and commodity-market consistency checks. - Missing components are explicitly marked as unavailable or block report generation; the workflow never inserts a default neutral value.
- Snapshots and bundles are audit evidence and must not be manually edited. Recollect with
capture, then rerunprepareandvalidate.
Plan & Review Advanced Features
- Three-stage lifecycle orchestration connects pre-market planning, midday validation, and post-market review while allowing independent midday and post-market reports when prerequisites are satisfied.
- Hard data-quality gates stop generation for missing, expired, inconsistent, or insufficiently covered inputs.
- Evidence-first reporting requires inline
[E##]citations for quantitative claims and checks bundle IDs, dates, and evidence numbering. - Versioned derived metrics provide reproducible Fear & Greed calculations through
derived_metrics.pyandfg-v1. - Structured hypothesis testing distinguishes confirmed, partially confirmed, invalidated, and insufficient-evidence outcomes.
- Six-stage theme lifecycle taxonomy separates observation, launch, expansion, climax, divergence, and retreat from independent dimensions such as breadth and fund-flow consistency.
- Snapshot comparison uses original midday and closing data instead of reverse-engineering values from earlier report text.
- Template enforcement restricts formal reports to approved sections and fields, excluding tool execution details and unsupported conclusions.
- The skill is read-only and research-focused: it collects data, creates plans, journals, and review conclusions without trade execution or delegated execution tools.
- Openclaw Skills users gain an auditable, repeatable workflow for market analysis and trading discipline review.
SKILL.md
Loading
Related Openclaw Skills
Sync Config with Upstream Sample
Safely refreshes reference comments and URLs from an upstream sample without changing the user's live configuration.
SOK Jobs Board
Fetches the SOK jobs board and clearly summarizes available Solana agent work, rewards, status, and delivery requirements.
Autopilot Guiones F100K
Autopilot Guiones F100K finds the day’s most relevant AI news and turns it into three recording-ready short-form video scripts.
Calendarizador de Contenido FÓRMULA 100K
Convierte guiones, ideas o referencias virales en un calendario de contenido semanal listo para producir, con exportación a CSV y XLSX.
Baidu Unlimited-OCR Document Parser
Baidu Unlimited-OCR Document Parser converts complex PDFs, office files, and images into structured Markdown without model deployment.
PDF Creation with fpdf2
Create lightweight, searchable, Unicode-safe PDF documents with Python and fpdf2 when reportlab is unavailable or unsuitable.
Featured*
