AI agent skill · by @SpaceZephyr

Baokuan Article Analysis

Fetch and analyze WeChat Official Account hot articles / 公众号爆款文章 by sector or keywords. Use when the user asks for 爆款文章分析, 赛道爆款, 公众号爆款数据, reading counts, likes, shares, comments, title patterns, writing style, 爆款原因分析, or writing references for content creation.

Archived 2026-10-23 · Source updated 2026-08-26. Check the original repository for the latest version.

SKILL.md instructions

# 爆款文章分析

## Overview

Use this skill to fetch hot WeChat Official Account article data by sector and generate a daily analysis report. It is for sector-level or keyword-level analysis, not exact historical scraping for one specific account.

The bundled script queries the hot-article data source, merges keywords by sector, deduplicates articles, ranks them, and writes:

- `data.json`: raw structured data
- `report.html`: a minimal visual analysis report with KPI cards, bar charts, ranked article cards, writing style analysis, hot reasons, and writing references

Do not default to Markdown reports. The primary user-facing artifact is `report.html`.

## Quick Start

Run with the default sectors:

```bash
python3 ~/.codex/skills/baokuan-article-analysis/scripts/daily_sector_trends.py \
  --output-dir ./output/baokuan-article-analysis
```

Run custom sectors:

```bash
python3 ~/.codex/skills/baokuan-article-analysis/scripts/daily_sector_trends.py \
  --sector 'AI Agent=AI Agent,智能体,Agent框架' \
  --sector 'Skill=skill,Skills,AI Skill' \
  --sector 'Claude Code=Claude Code,Codex,AI编程' \
  --output-dir ./output/baokuan-article-analysis
```

Run with a JSON config:

```bash
python3 ~/.codex/skills/baokuan-article-analysis/scripts/daily_sector_trends.py \
  --sector-config ~/.codex/skills/baokuan-article-analysis/references/default-sectors.json \
  --days 7 \
  --output-dir ./output/baokuan-article-analysis
```

## Workflow

1. Identify sectors and keywords from the user request.
2. If the user only gives broad sectors, use or adapt `references/default-sectors.json`.
3. Run `scripts/daily_sector_trends.py`.
4. Open the generated `report.html`.
5. Summarize for the user:
   - highest-reading and highest-sharing articles
   - writing style patterns
   - hot article reasons
   - title and topic formulas
   - practical writing references
6. Return the HTML file path as the main artifact.

## Script Options

| Option | Purpose |
|---|---|
| `--sector '赛道=关键词1,关键词2'` | Add one sector. Can repeat. |
| `--sector-config path.json` | Load sectors from JSON object. |
| `--days N` | Lookback window. Default is 7 days. |
| `--start-date YYYY-MM-DD` | Explicit start date. Overrides `--days`. |
| `--max-items-per-sector N` | Limit ranked articles per sector. Default is 10. |
| `--output-dir DIR` | Parent output directory. A date folder is created inside. |
| `--report-date YYYY-MM-DD` | Report date. Default is today. |

## Data Boundaries

- This skill returns hot-list data, not a specific account’s complete recent history.
- `clicksCount` is a public data-source snapshot and may lag behind live WeChat backend reads.
- If a specific account name returns no data, switch to that account’s topic keywords and compare same-sector articles.
- If today’s data is sparse, use `--days 7` or `--days 30`.

## Output Interpretation

Use reading count for reach, share count for spread, comments for discussion, and low-fan high-reading entries for title/structure references. Repeated accounts indicate strong competitors or content sources worth following.