uptrend-analyzer

Uptrend Analyzer Skill

Purpose

Diagnose market breadth health using Monty's Uptrend Ratio Dashboard, which tracks ~2,800 US stocks across 11 sectors. Generates a 0-100 composite score (higher = healthier) with exposure guidance.

Unlike the Market Top Detector (API-based risk scorer), this skill uses free CSV data to assess "participation breadth" - whether the market's advance is broad or narrow.

When to Use This Skill

English:

• User asks "Is the market breadth healthy?" or "How broad is the rally?"
• User wants to assess uptrend ratios across sectors
• User asks about market participation or breadth conditions
• User needs exposure guidance based on breadth analysis
• User references Monty's Uptrend Dashboard or uptrend ratios

Japanese:

• 「市場のブレドスは健全？」「上昇の裾野は広い？」
• セクター別のアップトレンド比率を確認したい
• 相場参加率・ブレドス状況を診断したい
• ブレドス分析に基づくエクスポージャーガイダンスが欲しい
• Montyのアップトレンドダッシュボードについて質問

Prerequisites

• Python 3.9+ with the requests library (CSV parsing uses the stdlib csv/io modules)
• Internet connection to fetch CSV data from GitHub (no API key required)
• No paid API subscriptions needed

Difference from Market Top Detector

Execution Workflow

Phase 1: Execute Python Script

Run the analysis script (no API key needed):

python3 skills/uptrend-analyzer/scripts/uptrend_analyzer.py

The script will:

1. Download CSV data from Monty's GitHub repository
2. Calculate 5 component scores
3. Generate composite score and reports

Phase 2: Present Results

Present the generated Markdown report to the user, highlighting:

• Composite score and zone classification
• Exposure guidance (Full/Normal/Reduced/Defensive/Preservation)
• Sector heatmap showing strongest and weakest sectors
• Key momentum and rotation signals

5-Component Scoring System

Scoring Zones

7-Level Zone Detail

Each scoring zone is further divided into sub-zones for finer-grained assessment:

Warning System

Active warnings trigger exposure penalties that tighten guidance even when the composite score is high:

Penalties stack (max -10) + multi-warning discount (+1 when ≥2 active). Applied after composite scoring.

Momentum Smoothing

Slope values are smoothed using EMA(3) (Exponential Moving Average, span=3) before scoring. Acceleration is calculated by comparing the recent 10-point average vs prior 10-point average of smoothed slopes (10v10 window), with fallback to 5v5 when fewer than 20 data points are available.

Historical Confidence Indicator

The Historical Context component includes a confidence assessment based on:

• Sample size: Number of historical data points available
• Regime coverage: Proportion of distinct market regimes (bull/bear/neutral) observed
• Recency: How recent the latest data point is

Confidence levels: High, Medium, Low.

API Requirements

Required: None (uses free GitHub CSV data)

Output Files

• JSON: uptrend_analysis_YYYY-MM-DD_HHMMSS.json
• Markdown: uptrend_analysis_YYYY-MM-DD_HHMMSS.md

Reference Documents

references/uptrend_methodology.md

• Uptrend Ratio definition and thresholds
• 5-component scoring methodology
• Sector classification (Cyclical/Defensive/Commodity)
• Historical calibration notes

When to Load References

• First use: Load uptrend_methodology.md for full framework understanding
• Regular execution: References not needed - script handles scoring

Same-run provenance

Pass --run-id daily-20261017-unique to both required market-regime-daily producers. Without it each producer generates its own unique ID. JSON includes schema 1.0 data_provenance with actual primary timeseries business date, UTC retrieval time, source and run ID. Unknown adjustment, point-in-time, survivorship and corporate-action flags are false. Auxiliary summary dates are unverified; this does not certify the freshness of every composite input. Missing or malformed primary dates stop report generation.

Pass the exact JSON paths printed by these invocations to Exposure Coach's --require-provenance --run-id … --as-of … mode. Run IDs are caller-provided consistency markers, not authenticity or point-in-time proof.