stock-review-ai（股票评论 AI）

🚀 Stock Review

👉 Live Demo Blog

GitHub: 👉 https://github.com/Donvink/stock-review

Language

Match user's language: Respond in the same language the user uses. If the user writes in Chinese, respond in Chinese. If the user writes in English, respond in English.

Script Directory

Agent Execution: Determine this SKILL.md directory as {baseDir}, then use {baseDir}/scripts/<name>.py. Ensure Python 3.10+ is installed and dependencies are configured.

Configuration Preferences

1. Check if config.yaml exists: {baseDir}/stock-review/config.yaml

1. Check if .env file exists and is configured with GEMINI_API_KEY, WECHAT_APP_ID, WECHAT_APP_SECRET: {baseDir}/stock-review/.env

config.yaml supports: Default publishing platforms | Whether to skip AI analysis by default | Default data backtracking days | Default request delay | Default retry count | API key configuration .env supports: API key configuration

Minimum supported keys (case-insensitive, accepts 1/0 or true/false):

Recommended config.yaml example:

# default configuration for stock review skill
review:
  markets:                          # can include "shanghai", "shenzhen", "hongkong"
    - "shanghai"
    - "shenzhen"
    - "hongkong"
  default_period: "daily"           # can be "daily", "weekly", "monthly"
  date: null                        # can be specific date "YYYYMMDD" like "20260101" or null for today
  force_refresh: false              # whether to force refresh data even if cached data is available
  skip_ai_analysis: false           # whether to skip AI analysis and just return raw data
  platforms: ["hugo"]               # platforms to publish the report, e.g. ['hugo', 'wechat'] or ['hugo'] or ['wechat']

paths:
  data_dir: null                    # directory to store fetched data and cache, null means current project directory

parameters:
  max_retries: 3
  request_delay: 0.5
  backtrack_days: 0
  
models:
  type: "gemini"
  model_name: "gemini-2.5-flash"

.env example:

# Gemini API Key
GEMINI_API_KEY="your_gemini_api_key"

# WeChat Official Account Configuration
WECHAT_APP_ID="your_wechat_app_id"
WECHAT_APP_SECRET="your_wechat_app_secret"

How to Get a Gemini API Key:

1. Visit the official portal: Go to https://aistudio.google.com/ and log in with your Google account.

1. Create an API Key: Click "Get API key" in the left sidebar, click "Create API key in new project", and copy the generated string (please save it securely—you won't be able to see the full key again after closing the window).

1. Important Notes:

Free Tier: Provides free quota but with request frequency limits (RPM/RPD). Data Privacy: Free tier data may be used for model improvement. For commercially sensitive data, consider enabling the paid mode.

How to Get WeChat Official Account Credentials:

1. Visit https://developers.weixin.qq.com/platform/
2. Navigate: My Business → Official Account → Development Keys
3. Add a development key, copy the AppID and AppSecret
4. Add the IP address of your machine to the whitelist

Environment Check

Before first use, install the dependencies.

pip install -r {baseDir}/requirements.txt

Check items: Python version | Dependencies | API keys | Network connection | Directory permissions

If any check fails, provide fix guidance:

Workflow Overview

Copy this checklist and check items as you progress:

Review Analysis Progress:
- [ ] Step 0: Load preferences (config.yaml, .env), determine execution parameters
- [ ] Step 1: Fetch market data
- [ ] Step 2: Run AI analysis (optional)
- [ ] Step 3: Generate report
- [ ] Step 4: Publish to platforms
- [ ] Step 5: Report complete

Step 0: Load Preferences

Check and load config.yaml settings (see Configuration Preferences section above), parse and store default values for subsequent steps.

Step 1: Fetch Market Data

Fetch the following data for the specified date:

Retry Mechanism:

• Default 3 retries
• 0.5 second request interval
• Automatic fallback to alternative interfaces on failure

Step 2: Run AI Analysis

CRITICAL: Run AI analysis only if:

• --skip-ai is not set
• GEMINI_API_KEY is configured (via config.yaml or environment variables)

AI Analysis Prompt:

prompt = f"""
Role Setting: You are a seasoned A-share strategy analyst with 20 years of experience...

Task Description: Conduct a multi-dimensional review based on the [daily review data]:
1. 🚩 Market Sentiment Diagnosis
2. 💰 Core Themes and Capital Flow
3. 🪜 Consecutive Limit-up Gradient and Space Game
4. ⚡ Key Stocks with Abnormal Movements Analysis
5. 🧭 Next Trading Day Strategy Recommendations

📊 Daily Review Data:
{market_summary}
"""

Output: data/{date}/ai_analysis_{date}.md

Step 3: Generate Reports

Market Summary Report:

• File: data/{date}/market_summary_{date}.md
• Format: Markdown
• Content: Tabular summary of all data

AI Analysis Report (if run):

• File: data/{date}/ai_analysis_{date}.md
• Format: Markdown
• Content: In-depth analysis generated by Gemini

Step 4: Publish to Platforms

Hugo Blog Publishing:

python3 {baseDir}/scripts/post_to_hugo.py --market-summary <file> --ai-analysis <file> --date <date>

Output: content/posts/stock-analysis-{YYYY-MM-DD}.md

WeChat Official Account Publishing (requires API credentials):

python3 {baseDir}/scripts/post_to_wechat.py --market-summary-file <file> --ai-analysis-file <file> --date <date> --cover-file <file> --title <title>

WeChat Official Account API Request Rules:

• Endpoint: POST https://api.weixin.qq.com/cgi-bin/draft/add?access_token=ACCESS_TOKEN
• article_type: news
• Requires thumb_media_id (cover image)
• Comment settings: need_open_comment=1, only_fans_can_comment=0

Step 5: Completion Report

Success Report:

✅ A-share Review Analysis Complete!

Date: 2026-03-04
Data: data/20260304/ (12 files)
AI Analysis: ✓ Generated (Gemini 2.0 Flash)

Published Platforms:
→ Hugo Blog: content/posts/stock-analysis-2026-03-04.md
→ WeChat Official Account: Draft ID: abc123def456

Market Snapshot:
• Shanghai Composite: 3350.52 (+1.02%)
• Turnover: 1.95 trillion
• Advance/Decline: 2857 / 2058
• Limit-up/Limit-down: 78 / 3

View Blog: https://donvink.github.io/stock-review/

Error Report:

❌ Review Analysis Failed

Error: Unable to fetch limit-up pool data
Suggestions: 
1. Check network connection
2. Try --force parameter to force refresh
3. Use --date to specify another date

Detailed Feature Description

Data Fetching Module

AI Analysis Module

Model: gemini-2.5-flash

Analysis Dimensions:

1. Market Sentiment Diagnosis - Advance/decline ratio, limit-up/down comparison, turnover
2. Core Theme Tracking - Capital flow, hot sectors
3. Consecutive Limit-up Gradient Analysis - Space board height, limit-up structure
4. Abnormal Movement Stock Analysis - High turnover, top traders list
5. Next Day Strategy Recommendations - Data-based trading suggestions

Publishing Module

Feature Comparison

Prerequisites

Required:

• Python 3.10+
• Dependencies: pip install -r requirements.txt
• Gemini API key (for AI analysis)

Optional:

• WeChat Official Account AppID and AppSecret (for WeChat publishing)
• Hugo blog environment (for blog publishing)

Configuration Locations (priority order):

1. CLI parameters
2. config.yaml, .env (project-level/user-level)
3. Environment variables
4. Default values

Troubleshooting

Extension Support

Customize via config.yaml. See the Configuration Preferences section for supported options.

Related References

Version History