Agent Skills › leamsigc/ShortsGenerator › ig-research

ig-research

GitHub

社交媒体内容研究工具，支持跨平台抓取、转录与分析。通过配置化架构实现多平台扩展，自动化执行从数据收集到报告生成的全流程，辅助进行内容策略与心理学分析。

skills/ig-research/SKILL.md leamsigc/ShortsGenerator

Trigger Scenarios

需要分析社交媒体热门内容执行跨平台数据抓取任务生成社交媒体分析报告

Install

npx skills add leamsigc/ShortsGenerator --skill ig-research -g -y

More Options

Use without installing

npx skills use leamsigc/ShortsGenerator@ig-research

指定 Agent (Claude Code)

npx skills add leamsigc/ShortsGenerator --skill ig-research -a claude-code -g -y

安装 repo 全部 skill

npx skills add leamsigc/ShortsGenerator --all -g -y

预览 repo 内 skill

npx skills add leamsigc/ShortsGenerator --list

SKILL.md

Frontmatter

{
    "name": "ig-research",
    "states": [
        "ig-research-setup    — Prerequisites, directory, config",
        "ig-research-scrape   — Scrape posts via Chrome DevTools",
        "ig-research-transcribe — Transcribe audio via Whisper",
        "ig-research-report   — Generate HTML report",
        "ig-research-analyze  — Content psychology analysis"
    ],
    "version": "1.0.0",
    "description": "Social media content research tool — scrape, transcribe, analyze top-performing\ncontent across platforms (Instagram, Facebook, LinkedIn, X, YouTube).\nMain orchestrator skill. Delegates to state sub-skills for each phase.\n",
    "allowed-tools": [
        "Bash",
        "Read",
        "Write",
        "Edit",
        "Glob",
        "Grep",
        "Question"
    ]
}

Social Research Tool — Main Skill

Full workflow: Setup → Scrape → Transcribe → Report → Analyze

Architecture

skills/ig-research/            # Skill (bundled, global-installable via ~/.claude/skills/)
├── SKILL.md
└── scripts/                   # Bundled — no external .ig-research/scripts needed
    ├── package.json
    ├── scrape.js
    ├── transcribe.sh
    └── report-html.js

.ig-research/                  # Tool data directory (project root, auto-resolved)
└── projects/<name>/
    ├── config.json            # Global per-project config
    ├── latest.json            # -> <YYYY-MM-DD_HHMMSS>
    └── <YYYY-MM-DD_HHMMSS>/   # One folder per run (dated)
        ├── raw-posts.json
        ├── transcripts/
        ├── hook-screenshots/
        └── report.html

Platform Extensibility

Each platform has its own scraper. The config schema includes a platform field:

Platform	Scraper location	Config `platform`
Instagram	`scripts/scrape.js` (default)	`instagram`
Facebook	`platforms/facebook/scrape.js`	`facebook`
LinkedIn	`platforms/linkedin/scrape.js`	`linkedin`
X	`platforms/x/scrape.js`	`x`
YouTube	`platforms/youtube/scrape.js`	`youtube`

To add a new platform: create a scraper script in skills/ig-research/platforms/<name>/scrape.js (or external platforms//scrape.js with --data-root) that outputs raw-posts.json with the same schema (posts array with postId, type, likes, caption, author, platform, hasScreenshots, hasAudio). The transcribe and report steps work with any platform's output.

Workflow

Load the full workflow with: skill ig-research

Each state can be run independently with: skill ig-research-<state>

# Run full pipeline for a project
skill ig-research-setup     # Step 1: Setup (prerequisites + config)
skill ig-research-scrape    # Step 2: Scrape Instagram
skill ig-research-transcribe # Step 3: Transcribe audio
skill ig-research-report    # Step 4: Generate HTML report
skill ig-research-analyze   # Step 5: Content psychology analysis