Are you the author? Sign in to claim
AI-driven global industrial policy tracking skill for Claude Code.
🇨🇳 中文 | 🇬🇧 English
An AI-powered global industrial policy tracking skill for Claude Code. Generates daily, weekly, and preview reports covering China/US/EU/Japan/Korea policies across semiconductors, AI, new energy, and supply chains — output as rich, visual HTML with expert-grade analysis.
一个 AI 驱动的全球产业政策追踪 Claude Code 技能。生成日报/周报/前瞻，覆盖中美欧日韩政策、半导体/AI/新能源/产业链，输出带深度专家拆解的可视化 HTML。
| 类型 | 触发词 | 频率 | 说明 |
|---|---|---|---|
| 周报 | 周报 / 政策周报 / weekly | 每周一 | 完整政策追踪：中国 5 大产业 + 国际 5 大领域 + 深度拆解 + 下周关注 |
| 日报 | 日报 / daily | 每日 | 当日宏观日报：重大事件 + 资产价格 + 科技突破扫描 + 专家拆解 |
| 前瞻 | 前瞻 / preview / 下周 | 每周日 | 下周前瞻：事件日历 + 核心数据预期 + 决策矩阵 |
The problem: Tracking global industrial policy is a high-effort, daily-grind task. A single report requires:
| Step | Time (manual) |
|---|---|
| Searching 8+ channels across CN/US/EU/JP/KR | 45–90 min |
| Reading and filtering relevant policies | 30–60 min |
| Writing expert-level analysis with historical context | 60–120 min |
| Formatting into a readable, visual report | 30–60 min |
| Total per daily report | ~3–5 hours |
| Total per weekly report | ~6–10 hours |
The fix: This skill automates the entire pipeline — multi-channel search, relevance filtering, expert analysis with dissenting viewpoints, and rich HTML output. A report that took half a day now takes one Claude Code invocation.
核心痛点： 全球产业政策追踪是高频、高信息密度的重复性工作。一个人手动做需要 3-10 小时/份。这个 skill 把搜索→过滤→分析→排版全流程自动化，将半天的工作压缩到一次 Claude Code 调用。
# Clone the repository
git clone https://github.com/oyj123321/claude-code-policy-tracker.git
# Link into Claude Code skills — project-level (current project only)
mkdir -p .claude/skills
# Linux / macOS
ln -s $(pwd)/claude-code-policy-tracker .claude/skills/policy-tracker
# Windows (PowerShell, as Administrator)
New-Item -ItemType SymbolicLink -Path .claude/skills/policy-tracker -Target (Resolve-Path claude-code-policy-tracker)
# Or: user-level (all projects)
# Linux / macOS
ln -s $(pwd)/claude-code-policy-tracker ~/.claude/skills/policy-tracker
npx openskills install oyj123321/claude-code-policy-tracker
cp -r claude-code-policy-tracker .claude/skills/policy-tracker
All three report types are invoked via the same skill:
/policy-tracker 日报
/policy-tracker 周报
/policy-tracker 前瞻
Or use natural language triggers:
The skill will:
D:\news\The pipeline runs two phases in parallel:
Phase 1: Chinese Platform Intelligence (Agent-Reach, parallel 4-way)
Many first-hand Chinese AI/industry insights surface on social platforms, not English media. Using Agent-Reach:
Agent-Reach config:
~/.agent-reach/config.yaml. Xiaohongshu/Weibo require Cookie (say "帮我配小红书" to Claude to auto-configure). Use a dedicated alt account.
Phase 2: English Web Search (WebSearch, 4+4 rounds)
Round 1 (parallel 4-way):
Round 2 (on-demand 3-4-way): 5. 中国专项债发行进度 + 以旧换新补贴具体金额 6. EU-China EV tariff / anti-subsidy / price undertaking latest 7. 台积电/三星/SK 海力士海外工厂进展 8. 下周全球经济数据日历（前瞻专用）
claude-code-policy-tracker/
├── SKILL.md # Core skill file · 核心技能文件 (Claude Code entry point)
├── README.md # This document · 本文档 (bilingual human-readable docs)
├── CHANGELOG.md # Version history · 版本历史
├── LICENSE # MIT License · MIT 许可
└── .gitignore # Git ignore rules
This skill is opinionated but extensible. If you have ideas for improvement:
Contributions are welcome under the MIT license.
MIT — see LICENSE for full text.
随意使用、修改、分发。Use freely, modify, distribute.
WCAG accessibility audit — automated scanning, manual review, remediation
Playwright testing toolkit — test generation, flaky test fix, migration helpers
Parallel development with git worktrees — port isolation, env sync, cleanup
1000+ skills curated from Anthropic, Vercel, Stripe, and other engineering teams