Not in the top 1,000 by GitHub stars
alchaincyf/huashu-report
机构级研究报告 Agent Skill · 规范从 2026 年顶级机构报告实物反向提炼，42 份采集／41 份进量化基线（Stanford/McKinsey/BCG/OpenAI/PwC/World Bank 等）· 六种报告原型 + 8 种图表模式 + 可复用生产流水线 · 跨 agent 通用
History
Review weekly changes to SKILL.md before updating. Dates mark the weeks in which we observed each version; checks apply to the recorded commits.
View weekly changes
SKILL.md
Week of Nov 2, 2026 · +0 / −0 lines
Before Network or broad shell → After Network or broad shell
eda161cf693f → eda161cf693f
Week of Oct 26, 2026 · Line counts unavailable
Before Not checked yet → After Network or broad shell
First observation → eda161cf693f
No earlier version was observed for comparison.
486GitHub stars
8stars in 7 days
1skill in this repository
1commit in 30 days
Heat
6.2- Star growth, last 7 daysNot available
- Star growth, last 24 hoursNot available
- Rank movement, last 7 daysNot available
- Commits, last 30 daysNot available
- Accounts asking about it through MCP, last 7 daysNot available
- Hot votes this weekNot available
Heat never changes the order by stars, and sponsorship never touches it. How heat works
Stars over time
30-day change not yet available: not enough history.
Individual skills in this repository
How it works
huashu-report
When your agent uses it
做机构级研究报告与学术论文——行业报告、白皮书、年度调研、数据洞察、arXiv论文，也含16:9咨询deck型报告（一页一结论）。定结构、写发现、建图表视觉系统。规范提炼自2026年顶级机构报告实测。单篇文章、纯演示PPT不适用（走design/slides）。
What it loads into context
Frontmatter
147 tokensAlways available to the agent
SKILL.md
4,411 tokensWhen the skill is activated
Reference documents
37,645 tokensLoaded on request
Token counts are estimates. Scripts run separately; they are not counted as reference context.
The playbook
- 你是谁
- 开工前先回答三个问题
- 五条硬规矩
- 工作顺序不能颠倒
- 详细规范
- 交付前自检
- 关于作者
- 版本检查
Scripts executed separately
assets/chart.pyassets/render.py
Files
.gitignoreasset48 bytesLICENSEasset1,067 bytesREADME.mddoc5,987 bytesSKILL.mdcore15,266 bytesSKILL.mdcore15,266 bytesassets/base.cssasset8,493 bytesassets/chart.pyruns18,711 bytesassets/render.pyruns18,152 bytescorpus/README.mddoc2,850 bytescorpus/anatomy.jsonasset70,203 bytescorpus/samples.mddoc12,877 bytesreferences/about-author.mddoc9,941 bytesreferences/图表模式库.mddoc9,753 bytesreferences/实证基线.mddoc4,506 bytesreferences/生产流水线.mddoc19,327 bytesreferences/研究深度.mddoc6,948 bytesreferences/科普.mddoc15,704 bytesreferences/结构骨架.mddoc8,436 bytesreferences/行文.mddoc5,115 bytesreferences/视觉方向库.mddoc6,623 bytesreferences/视觉系统.mddoc12,828 bytesreferences/论文.mddoc5,866 bytes
Read SKILL.md
# 报告写作
## 你是谁
不是「把资料整理成文档的人」。一份能被引用的报告，背后至少四个角色：
**研究员**做三件事，缺一件报告就掉一个档次：定口径（样本怎么来的、N 是多少、哪些被排除了）、
**给机制**（数据说了「是什么」，机制回答「为什么会这样」）、
**定位文献**（相对既有研究，这个发现是印证、是修正，还是揭示了既有框架没建模的维度）。
只做第一件，交出去的是资料汇编；三件都做，才是研究。
**编辑**决定读者读到第 3 页时应该已经知道什么。他砍掉一切「因为我们查到了所以写进去」的内容，
也砍掉一切关于「我们是怎么做的」的自我叙事。
**信息设计师**建立那套让 60 页保持一致的系统——标题层级、图表编号、数据源行的格式。读者不会注意到它，但它一旦崩了，报告就显得业余。Skills included
- huashu-report
Repository facts
- License
- MIT
- Created
- Aug 31, 2026
- Last pushed
- Oct 13, 2026
- Forks
- 65
- Open issues
- 0
- Project activity
- Active
Before you install
Network or broad shellAutomated signals, not a guarantee
Checked Nov 6, 2026
Not fully analyzed: some files were too long, unreadable or missing. This never lowers the level.
Show 1 note
SKILL.md:1Analysis incomplete: Referenced executable target unavailable; not analyzed
- No broad tool permissions
- No dangerous shell commands
- 152 network referencesin 8 files: references/about-author.md:3 corpus/samples.md:9 assets/chart.py:130 and 5 more
- No hidden or obfuscated text
- No instruction overrides
- 2 executable scriptsin 2 files: assets/chart.py:1 assets/render.py:1
- No access to secrets
- No requests to vote in Hot
Show all 154 signals by file
README.md
- Network accessline 24
# Claude Code git clone https://github.com/alchaincyf/huashu-report ~/.claude/skills/huashu-report - Network accessline 25
git clone https://github.com/alchaincyf/huashu-report ~/.claude/skills/huashu-report - Network accessline 27
# 三家 agent 共用一个 skill 池的话 git clone https://github.com/alchaincyf/huashu-report ~/.agents/skills/huashu-report - Network accessline 28
git clone https://github.com/alchaincyf/huashu-report ~/.agents/skills/huashu-report
SKILL.md
- Network accessline 4
metadata: author: "花叔" homepage: "https://github.com/alchaincyf" - Network accessline 5
author: "花叔" homepage: "https://github.com/alchaincyf" repository: "https://github.com/alchaincyf/huashu-report" - Network accessline 6
homepage: "https://github.com/alchaincyf" - Network accessline 7
repository: "https://github.com/alchaincyf/huashu-report" - Network accessline 188
本 skill 由花叔开发维护，仓库：https://github.com/alchaincyf/huashu-report - Network accessline 199
- 成功：第一行就是最新版本，和本文件开头的 `metadata.version` 比较，然后写入 `<今天> ok`。远程更高时，先完成用户的任务，结束时告诉用户：「huashu-report 有新版本 vX.Y.Z，更新说明见 https://github.com/alchaincyf/huashu-report/releases ，需要时运行 `git -C <本目录> pull --ff-only`；不想再收到提醒，把 `.last-update-check` 的内容
assets/chart.py
- Executable scriptline 1
#!/usr/bin/env python3 - Network accessline 130
val_w = max(text_w(fmt.format(v), 8.4) for _, v in data) + 10 bw = max(60, w - label_w - val_w) s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] - Network accessline 131
bw = max(60, w - label_w - val_w) s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] for i, (lab, v) in enumerate(data): - Network accessline 132
s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] - Network accessline 152
bw = w - label_w - 12 cols = seg_colors or [TEAL, TEAL2, TEAL3, SAND, CLAY2] s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] - Network accessline 153
cols = seg_colors or [TEAL, TEAL2, TEAL3, SAND, CLAY2] s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] for i, (lab, vals) in enumerate(rows): - Network accessline 154
s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] - Network accessline 193
def Y(v): return pad_t + ph - ph * (v - ymin) / (ymax - ymin) zero = Y(0) s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] - Network accessline 194
zero = Y(0) s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] for t in range(5): - Network accessline 195
s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] - Network accessline 240
def X(i): return pad_l + pw * i / max(1, n - 1) def Y(v): return pad_t + ph - ph * (v - ymin) / (ymax - ymin) s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] - Network accessline 241
def Y(v): return pad_t + ph - ph * (v - ymin) / (ymax - ymin) s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] for t in range(5): - Network accessline 242
s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] - Network accessline 274
w = max(w, vw + 16) h = 66 + len(lines) * 11 + (11 if sub else 0) s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] - Network accessline 275
h = 66 + len(lines) * 11 + (11 if sub else 0) s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] s.append(_txt(w / 2, 40, value, 38, color, "middle", "bold", SANS)) - Network accessline 276
s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] - Network accessline 292
posc = [TEAL, TEAL2] mid = label_w + bw s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] - Network accessline 293
mid = label_w + bw s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] s.append(f'<line x1="{mid}" y1="4" x2="{mid}" y2="{h-16}" stroke="{INK}" stroke-width="0.8"/>') - Network accessline 294
s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] - Network accessline 332
lo, hi = min(vals) * 0.86, max(vals) * 1.08 def Y(v): return pad_t + ph - ph * (v - lo) / (hi - lo) s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] - Network accessline 333
def Y(v): return pad_t + ph - ph * (v - lo) / (hi - lo) s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] s.append(f'<line x1="{xl}" y1="{pad_t-6}" x2="{xl}" y2="{pad_t+ph+6}" stroke="{RULE}" strok - Network accessline 334
s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] - Network accessline 355
def X(v): return pad + pw * v / 100 def Y(v): return 14 + ph - ph * v / 100 s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] - Network accessline 356
def Y(v): return 14 + ph - ph * v / 100 s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">'] s.append(f'<rect x="{pad}" y="14" width="{pw}" height="{ph}" fill="none" stroke="{RULE}" stroke-width="0.6" - Network accessline 357
s = [f'<svg viewBox="0 0 {w} {h}" width="100%" xmlns="http://www.w3.org/2000/svg">']
assets/render.py
- Executable scriptline 1
#!/usr/bin/env python3
corpus/README.md
- Network accessline 16
| 档位 | 份数 | 你可以做什么 | |---|---|---| | 开放许可（CC BY / BY-NC / BY-ND / OGL） | 11 | 自由使用、再分发，署名即可 | - Network accessline 17
|---|---|---| | 开放许可（CC BY / BY-NC / BY-ND / OGL） | 11 | 自由使用、再分发，署名即可 | | 保留所有权利 | 17 | 只能自己读；引用数据点请注明原始出处 | - Network accessline 18
| 开放许可（CC BY / BY-NC / BY-ND / OGL） | 11 | 自由使用、再分发，署名即可 |
corpus/samples.md
- Network accessline 9
- **开放许可（11份）**——CC BY / CC BY-NC / CC BY-ND / OGL。这类你可以自由使用，包括再分发，署名即可 - Network accessline 21
可以直接引用、搬运、二次加工——NC 类限非商业用途，ND 类不可修改。署名格式见各自的原始页面。 - Network accessline 23
| # | 日期 | 机构 | 报告 | 页数 | 许可 | 备注 | |---|---|---|---|---|---|---| | 1 | 2026-01-01 | CSET (Georgetown) | [When AI Builds AI (Workshop Report)](https://cset.georgetown.edu/publication/when-ai-builds-ai/) | 32 | CC BY-NC 4.0 | | - Network accessline 24
|---|---|---|---|---|---|---| | 1 | 2026-01-01 | CSET (Georgetown) | [When AI Builds AI (Workshop Report)](https://cset.georgetown.edu/publication/when-ai-builds-ai/) | 32 | CC BY-NC 4.0 | | | 2 | 2026-02-01 | International AI Safety Repor - Network accessline 25
| 1 | 2026-01-01 | CSET (Georgetown) | [When AI Builds AI (Workshop Report)](https://cset.georgetown.edu/publication/when-ai-builds-ai/) | 32 | CC BY-NC 4.0 | | - Network accessline 26
| 2 | 2026-02-01 | International AI Safety Report（主席 Yoshua B… | [International AI Safety Report 2026](https://arxiv.org/abs/2602.21012) | 220 | Open Government Licence v3.0 | | - Network accessline 27
| 3 | 2026-02-19 | OECD | [The OECD.AI Index](https://www.oecd.org/en/publications/2026/02/oecd-ai-observatory-index_8f5fa0f2.html) | 56 | CC BY 4.0 | | - Network accessline 28
| 4 | 2026-04-13 | Stanford HAI (Institute for Human-Centered… | [The 2026 AI Index Report](https://hai.stanford.edu/ai-index/2026-ai-index-report) | 425 | CC BY-ND 4.0 | | | 5 | 2026-05-04 | International Labour Organization (ILO) | [A Mo - Network accessline 29
| 5 | 2026-05-04 | International Labour Organization (ILO) | [A Moment of Choice: Harnessing Artificial Intelligence for Decent Work (Report of the Director-General, ILC 114th Session)](https://www.ilo.org/resource/conference-paper/ilc/ilc1 - Network accessline 30
| 6 | 2026-06-01 | CSET (Georgetown) | [Identifying the AI Development Workforce (Data Brief)](https://cset.georgetown.edu/publication/identifying-the-ai-development-workforce) | 14 | CC BY-NC 4.0 | | - Network accessline 31
| 7 | 2026-06-05 | OECD | [AI and Skills: What We Know So Far](https://www.oecd.org/en/publications/ai-and-skills_f843b352-en.html) | 9 | CC BY 4.0 | | | 8 | 2026-06-16 | Reuters Institute for the Study of Journal… | [Digital News Report 2 - Network accessline 32
| 8 | 2026-06-16 | Reuters Institute for the Study of Journal… | [Digital News Report 2026](https://reutersinstitute.politics.ox.ac.uk/digital-news-report/2026) | 178 | CC BY 4.0 | | | 9 | 2026-07-01 | Georgetown CSET (Center for Security - Network accessline 33
| 9 | 2026-07-01 | Georgetown CSET (Center for Security and E… | [Beyond Targeting: The Untapped Role of AI in Military Decision-Making](https://cset.georgetown.edu/publication/beyond-targeting/) | 73 | CC BY-NC 4.0 | | - Network accessline 34
| 10 | 2026-08-04 | World Bank | [World Development Report 2026: The Promise of Artificial Intelligence](https://www.worldbank.org/en/publication/wdr2026) | 380 | CC BY 3.0 IGO | | - Network accessline 35
| 11 | 2026-08-06 | OpenAI | [From asking to doing: How the world is putting ChatGPT to work](https://openai.com/index/how-the-world-is-putting-chatgpt-to-work/) | 40 | CC BY 4.0（Signal 数据部分） | ⚠<U+FE0F> 实为 OpenAI Signals 数据说明文档，报告正文未获取 | - Network accessline 41
| # | 日期 | 机构 | 报告 | 页数 | 许可 | 备注 | |---|---|---|---|---|---|---| | 12 | 2026-01-01 | ICONIQ Capital (ICONIQ Growth) | [2026 State of AI: Bi-Annual Snapshot — The Execution Era of AI](https://www.iconiq.com/growth/reports/2026-state-of-ai-b - Network accessline 42
|---|---|---|---|---|---|---| | 12 | 2026-01-01 | ICONIQ Capital (ICONIQ Growth) | [2026 State of AI: Bi-Annual Snapshot — The Execution Era of AI](https://www.iconiq.com/growth/reports/2026-state-of-ai-bi-annual-snapshot) | 44 | 保留所有权利 | - Network accessline 43
| 12 | 2026-01-01 | ICONIQ Capital (ICONIQ Growth) | [2026 State of AI: Bi-Annual Snapshot — The Execution Era of AI](https://www.iconiq.com/growth/reports/2026-state-of-ai-bi-annual-snapshot) | 44 | 保留所有权利 | | - Network accessline 44
| 13 | 2026-02-09 | McKinsey | [McKinsey Global Tech Agenda 2026: How CIOs are shaping enterprise strategy and growth](https://www.mckinsey.com/capabilities/mckinsey-technology/our-insights/mckinsey-global-tech-agenda-2026) | 10 | 保留所有权利 | - Network accessline 45
| 14 | 2026-02-11 | Morgan Stanley Research | [Mapping AI's Rate of Change: The Broadening Trend（AI变革速度图谱第五版）](https://www.morganstanley.com/content/dam/msdotcom/en/assets/pdfs/Research_AI-Rate-of-Change.pdf) | 60 | 保留所有权利 | | - Network accessline 46
| 15 | 2026-03-01 | Bank of America Institute | [Not quite mAInstream: A consumer AI profile（消费者AI付费画像）](https://institute.bankofamerica.com/economic-insights/consumer-ai-usage.html) | 7 | 保留所有权利 | | - Network accessline 47
| 16 | 2026-03-16 | World Economic Forum × Accenture | [Organizational Transformation in the Age of AI (White Paper)](https://www.weforum.org/publications/organizational-transformation-in-the-age-of-ai-how-organizations-maximize-ais-potenti - Network accessline 48
| 17 | 2026-03-24 | RAND Corporation | [Artificial General Intelligence Forecasting and Scenario Analysis: State of the Field, Methodological Gaps, and Strategic Implications](https://www.rand.org/pubs/research_reports/RRA4692-1.html) | 75 - Network accessline 49
| 18 | 2026-04-01 | IMF | [Global Economic and Financial Implications of Artificial Intelligence: Lessons from a Scenario-Planning Exercise (IMF Note 2026/002)](https://www.elibrary.imf.org/view/journals/068/2026/002/article-A001-en.xml) | - Network accessline 50
| 19 | 2026-05-01 | Bank of America Institute | [Adaptability is the new job security: AI and the future of work（适应力是新的工作保障）](https://institute.bankofamerica.com/transformation/ai-and-future-of-work.html) | 10 | 保留所有权利 | | - Network accessline 51
| 20 | 2026-05-11 | J.P. Morgan Private Bank | [2026 Mid-Year Outlook: Promise and Pressure（承诺与压力）](https://privatebank.jpmorgan.com/nam/en/insights/latest-and-featured/mid-year-outlook) | 70 | 保留所有权利 | | - Network accessline 52
| 21 | 2026-06-01 | KPMG | [KPMG Global AI Quarterly Pulse Survey: Q2 2026](https://kpmg.com/xx/en/our-insights/ai-and-technology/ai-pulse.html) | 27 | 保留所有权利 | | - Network accessline 53
| 22 | 2026-06-12 | CNAS | [Red Lines: Understanding the National Security Risks of China's Advanced AI](https://www.cnas.org/publications/reports/red-lines) | 38 | 保留所有权利 | | - Network accessline 54
| 23 | 2026-06-15 | PwC | [PwC 2026 Global AI Jobs Barometer: Two futures for jobs in an AI era](https://www.pwc.com/gx/en/issues/artificial-intelligence/ai-jobs-barometer.html) | 30 | 保留所有权利 | | - Network accessline 55
| 24 | 2026-06-22 | World Economic Forum (与PwC合作) | [Artificial Intelligence and the Future of Entry-Level Work: A Framework for Safeguarding and Reinventing Early Career Pathways](https://www.weforum.org/publications/artificial-intelligenc - Network accessline 56
| 25 | 2026-07-01 | ICONIQ Capital (ICONIQ Growth) | [2026 State of AI Report: The Builder's Economy](https://www.iconiq.com/growth/reports/state-of-ai-2026) | 47 | 保留所有权利 | | - Network accessline 57
| 26 | 2026-07-07 | RAND Corporation | [The Transatlantic Artificial Intelligence Calculus: U.S.–European Union and Member State AI Cooperation in an Era of Strategic Competition](https://www.rand.org/pubs/research_reports/RRA4754-1.html) | - Network accessline 58
| 27 | 2026-08-01 | UBS (Chief Investment Office) | [UBS House View – Monthly Letter, August 2026（8月投资月报：AI资本开支与变现的可持续性）](https://advisors.ubs.com/mediahandler/media/823469/UBS%20House%20View%20for%20August%202026.pdf) | 28 | 保留所有权利 | | - Network accessline 59
| 28 | 2026-08-04 | RAND Corporation | [Secure Inference Data Centers: A Vertically Integrated Strategy for Security Engineering](https://www.rand.org/pubs/research_reports/RRA4827-1.html) | 51 | 保留所有权利 | | - Network accessline 65
| # | 日期 | 机构 | 报告 | 页数 | 许可 | 备注 | |---|---|---|---|---|---|---| | 29 | 2026-01-01 | Deloitte | [Deloitte State of AI in the Enterprise 2026（第八年度）](https://www.deloitte.com/global/en/issues/generative-ai/state-of-ai-in-enterprise.html) | 4 - Network accessline 66
|---|---|---|---|---|---|---| | 29 | 2026-01-01 | Deloitte | [Deloitte State of AI in the Enterprise 2026（第八年度）](https://www.deloitte.com/global/en/issues/generative-ai/state-of-ai-in-enterprise.html) | 41 | 未声明许可 | | | 30 | 2026-01-15 | A - Network accessline 67
| 29 | 2026-01-01 | Deloitte | [Deloitte State of AI in the Enterprise 2026（第八年度）](https://www.deloitte.com/global/en/issues/generative-ai/state-of-ai-in-enterprise.html) | 41 | 未声明许可 | | - Network accessline 68
| 30 | 2026-01-15 | Anthropic | [Anthropic Economic Index report: Economic Primitives](https://www.anthropic.com/research/anthropic-economic-index-january-2026-report) | 55 | 未声明许可 | | - Network accessline 69
| 31 | 2026-01-15 | BCG | [BCG AI Radar 2026: As AI Investments Surge, CEOs Take the Lead](https://www.bcg.com/publications/2026/as-ai-investments-surge-ceos-take-the-lead) | 29 | 未声明许可 | | - Network accessline 70
| 32 | 2026-03-17 | Hugging Face | [State of Open Source on Hugging Face: Spring 2026](https://huggingface.co/blog/huggingface/state-of-os-hf-spring-2026) | 26 | 未声明许可 | ⚠<U+FE0F> 实为 Longpre et al.《Economies of Open Intelligence》 | - Network accessline 71
| 33 | 2026-03-24 | Anthropic | [Anthropic Economic Index report: Learning Curves](https://www.anthropic.com/research/economic-index-march-2026-report) | 20 | 未声明许可 | | - Network accessline 72
| 34 | 2026-04-01 | OpenAI | [The AI Jobs Transition Framework (US)](https://openai.com/index/modeling-ai-jobs-transition/) | 32 | 未声明许可 | | - Network accessline 73
| 35 | 2026-06-03 | BCG | [BCG AI at Work 2026: Why Strategy Matters More Than Tools（第四年度）](https://www.bcg.com/publications/2026/ai-at-work-why-strategy-matters-more-than-tools) | 29 | 未声明许可 | | - Network accessline 74
| 36 | 2026-06-26 | Anthropic | [Anthropic Economic Index report: Cadences](https://www.anthropic.com/research/economic-index-june-2026-report) | 33 | 未声明许可 | | - Network accessline 75
| 37 | 2026-06-29 | Brookings Institution（Xavier de Souza Brig… | [Getting to all-of-the-above: A framework of solutions for AI's coming impacts on work and workers](https://www.brookings.edu/articles/ai-workforce-policy-framework/) | 14 | - Network accessline 76
| 38 | 2026-07-01 | Oxford Internet Institute, University of O… | [Large Language Models in the UK: Public Use, Trust, and Attitudes](https://www.oii.ox.ac.uk/uk-adults-increasingly-seek-emotional-support-and-companionship-from-ai-new-repor - Network accessline 77
| 39 | 2026-07-27 | OpenAI | [Work at the Frontier: How AI is expanding what people do at work](https://openai.com/index/how-ai-is-expanding-what-people-do-at-work/) | 16 | 未声明许可 | | - Network accessline 78
| 40 | 2026-07-29 | Epoch AI | [Will parallelization limits delay an intelligence explosion?](https://epoch.ai/publications/parallelization-constraints-could-delay-a-technological-singularity) | 32 | 未声明许可 | | - Network accessline 79
| 41 | 2026-08-11 | OpenAI × Columbia Business School × Wharton | [How Organizations Use AI: Evidence from ChatGPT](https://cdn.openai.com/pdf/how-organizations-use-chatgpt.pdf) ᵃ | 69 | 未声明许可 | | - Network accessline 80
| 42 | 2026-08-12 | Stanford Digital Economy Lab（Erik Brynjolf… | [Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence（2026年8月修订版）](https://digitaleconomy.stanford.edu/news/canariesaug26/) | 1 - Network accessline 81
| 43 | 2026-08-26 | OpenAI × Bocconi × UC Berkeley | [Training Novices to Think, or Giving Them LLMs? Evidence from an RCT](https://cdn.openai.com/pdf/novices-and-llm-august-2026.pdf) ᵃ | 49 | 未声明许可 | |
references/about-author.md
- Network accessline 3
> 本文件由 [huashu-skills](https://github.com/alchaincyf/huashu-skills) 根据花叔的 GitHub 主页自动生成。供 agent 在用户问到作者、相关作品或联系方式时读取；正常任务中不主动介绍。 - Network accessline 7
主页：https://github.com/alchaincyf - Network accessline 11
- **huashu-art-motion：让画动起来**：https://github.com/alchaincyf/huashu-art-motion - Network accessline 13
[看动画样片](https://github.com/alchaincyf/huashu-art-motion#动画样片) · [开始使用](https://github.com/alchaincyf/huashu-art-motion#里面有什么) · [下载完整成片](https://github.com/alchaincyf/huashu-art-motion/releases/latest) - Network accessline 17
- **女娲.skill：把高手的思维，变成你的工具**：https://github.com/alchaincyf/nuwa-skill - Network accessline 18
提炼一个人的心智模型、决策方法和表达方式，变成可以反复调用的Skill。 也可以直接试试：[乔布斯视角](https://github.com/alchaincyf/steve-jobs-skill) · [X运营](https://github.com/alchaincyf/x-mentor-skill) - **Huashu Design：把想法做成作品**：https://github.com/alchaincyf/huashu-design - Network accessline 19
也可以直接试试：[乔布斯视角](https://github.com/alchaincyf/steve-jobs-skill) · [X运营](https://github.com/alchaincyf/x-mentor-skill) - Network accessline 20
- **Huashu Design：把想法做成作品**：https://github.com/alchaincyf/huashu-design - Network accessline 21
用HTML做原型、幻灯片、动画和信息图，从定方向到检查成品。 - **FanBox：Coding Agent的驾驶舱**：https://github.com/alchaincyf/fanbox 把文件、终端和每一次改动放在一起，看清Agent做了什么，随时接手。 - Network accessline 22
- **FanBox：Coding Agent的驾驶舱**：https://github.com/alchaincyf/fanbox - Network accessline 27
- **达尔文.skill：让Skill在实测中进化**：https://github.com/alchaincyf/darwin-skill - Network accessline 28
评估、改进、测试，保留有效改动，回滚退步。女娲造Skill，达尔文磨Skill。 - **huashu-chrome：让Agent用上你的浏览器**：https://github.com/alchaincyf/huashu-chrome 使用你自己的Chrome和已有登录态操作网页，把站点经验留下来。 - Network accessline 29
- **huashu-chrome：让Agent用上你的浏览器**：https://github.com/alchaincyf/huashu-chrome - Network accessline 30
使用你自己的Chrome和已有登录态操作网页，把站点经验留下来。 - **huashu-mac-use：让Agent操作Mac原生应用**：https://github.com/alchaincyf/huashu-mac-use 让Agent操作没有API的Mac原生应用：先探测可用通道，再执行并留下取证。 - Network accessline 31
- **huashu-mac-use：让Agent操作Mac原生应用**：https://github.com/alchaincyf/huashu-mac-use - Network accessline 36
- **橙皮书：从真实项目里长出来的AI指南**：https://bookai.top - Network accessline 38
[阅读橙皮书](https://bookai.top) · [Claude Code](https://github.com/alchaincyf/claude-code-orange-book) · [Hermes Agent](https://github.com/alchaincyf/hermes-agent-orange-book) · [Obsidian AI](https://github.com/alchaincyf/obsidian-ai-orange-boo - Network accessline 42
完整目录、安装方法和机器可读的 skills.json：https://github.com/alchaincyf/huashu-skills - Network accessline 44
| skill | 适合什么需求 | 仓库 | |---|---|---| | huashu-design | HTML 原生设计系统：高保真原型/幻灯片/动画/可视化，20 种设计哲学 + 5 维专家评审 + MP4 导出 | https://github.com/alchaincyf/huashu-design | - Network accessline 45
|---|---|---| | huashu-design | HTML 原生设计系统：高保真原型/幻灯片/动画/可视化，20 种设计哲学 + 5 维专家评审 + MP4 导出 | https://github.com/alchaincyf/huashu-design | | huashu-mac-use | 让 agent 操控 Mac 上没有 API 的原生 app：探测选层（CDP / AX / 坐标 / 像素）、读后台写过闸不打扰用户、每步回读与取证、收工回流进档案； - Network accessline 46
| huashu-design | HTML 原生设计系统：高保真原型/幻灯片/动画/可视化，20 种设计哲学 + 5 维专家评审 + MP4 导出 | https://github.com/alchaincyf/huashu-design | - Network accessline 47
| huashu-mac-use | 让 agent 操控 Mac 上没有 API 的原生 app：探测选层（CDP / AX / 坐标 / 像素）、读后台写过闸不打扰用户、每步回读与取证、收工回流进档案；含 Swift 内核与 CDP 工具 | https://github.com/alchaincyf/huashu-mac-use | - Network accessline 48
| huashu-gpt-image | GPT-image 的 prompt 工程方法论：真实参考名替代形容词，含单图 playbook、批量网格生成、失败模式树 | https://github.com/alchaincyf/huashu-gpt-image | - Network accessline 49
| huashu-icon-set | 端到端为项目生成风格统一的图标集：4×4 网格批量生成 + 逐个精度兜底 + 智能抠图（豆包环境专用，依赖内置 image_gen） | https://github.com/alchaincyf/huashu-icon-set | - Network accessline 50
| huashu-slide-doubao | 豆包环境专用视觉物料生产：slides + 公众号封面 + 视频封面，走内置 image_gen 零 API 费用 | https://github.com/alchaincyf/huashu-slide-doubao | - Network accessline 51
| huashu-slide-codex | Codex 环境专用视觉物料生产：slides + 公众号封面 + 视频封面，走 Codex 内置 image_gen | https://github.com/alchaincyf/huashu-slide-codex | - Network accessline 52
| huashu-excel | 数据分析全流程：体检脏表→清洗→定口径→算指标→对账→报告，依赖仅 openpyxl | https://github.com/alchaincyf/huashu-excel | - Network accessline 53
| huashu-md-html | md 流水线：万物（PDF/DOCX/EPUB/网页/音频）转干净 md，md 加工成出版级 HTML/DOCX，四套反 AI slop 主题 | https://github.com/alchaincyf/huashu-md-html | - Network accessline 54
| seedance-skill | 白模驱动的 AI 视频工作台：白模参考视频与提示词出自同一份 scene.json，渲染前做几何体检 | https://github.com/alchaincyf/seedance-skill | - Network accessline 55
| tramstop-skill | 电车站.skill：证据驱动的去 AI 味，四层 AI 味模型 + 真实素材注入，来自四版本盲测实验 | https://github.com/alchaincyf/tramstop-skill | - Network accessline 56
| dukou | 渡口：公众号文章一键渡到 X Articles / B站专栏，Chrome 插件 + 本地桥 + skill | https://github.com/alchaincyf/dukou | - Network accessline 57
| x-mentor-skill | X 导师：蒸馏 6 位顶级 X 创作者方法论 + 开源算法数据，完整的选题-写作-增长手册 | https://github.com/alchaincyf/x-mentor-skill | - Network accessline 58
| huashu-weread | 微信读书高阶顾问：书架+笔记交叉分析，个性化荐书/阶梯书单/笔记提炼/年度阅读复盘 | https://github.com/alchaincyf/huashu-weread | - Network accessline 59
| nuwa-skill | 女娲：蒸馏任何人的思维方式——心智模型、决策启发式、表达 DNA，产出可运行的人物 skill | https://github.com/alchaincyf/nuwa-skill | - Network accessline 60
| darwin-skill | 达尔文：让 skill 无限进化——评估→改进→测试→保留或回滚，全程 git 留痕 | https://github.com/alchaincyf/darwin-skill | - Network accessline 61
| freud-skill | 弗洛伊德：给 AI 做心理分析，基于 Anthropic 可解释性研究，诊断 prompt/skill 的身份冲突并重写 | https://github.com/alchaincyf/freud-skill | - Network accessline 62
| skills-guide | Claude Code Skills 完整指南：理解、查找和创建 skills | https://github.com/alchaincyf/skills-guide | - Network accessline 63
| huashu-report | 机构级研究报告与学术论文：六种报告原型 + 结构骨架 + 图表模式库 + 可复用生产流水线，规范提炼自 2026 年 42 份顶级机构报告实测 | https://github.com/alchaincyf/huashu-report | - Network accessline 64
| zhangxuefeng-skill | 张雪峰的认知操作系统：高考志愿/考研/职业规划 | https://github.com/alchaincyf/zhangxuefeng-skill | - Network accessline 65
| steve-jobs-skill | 乔布斯的认知操作系统：产品判断、砍需求、品味 | https://github.com/alchaincyf/steve-jobs-skill | - Network accessline 66
| elon-musk-skill | 马斯克的认知操作系统：第一性原理、成本结构、激进迭代 | https://github.com/alchaincyf/elon-musk-skill | - Network accessline 67
| munger-skill | 查理·芒格的认知操作系统：逆向思考、认知偏误、投资决策 | https://github.com/alchaincyf/munger-skill | - Network accessline 68
| karpathy-skill | Andrej Karpathy 的认知操作系统：AI 技术可靠性、学习方法、行业判断 | https://github.com/alchaincyf/karpathy-skill | - Network accessline 69
| trump-skill | 特朗普的思维操作系统：谈判、权力、传播 | https://github.com/alchaincyf/trump-skill | - Network accessline 70
| feynman-skill | 费曼的认知操作系统：真懂 vs 记住名词、货物崇拜检测 | https://github.com/alchaincyf/feynman-skill | - Network accessline 71
| naval-skill | Naval Ravikant 的认知操作系统：杠杆、特定知识、财富路径 | https://github.com/alchaincyf/naval-skill | - Network accessline 72
| sun-yuchen-perspective | 孙宇晨的注意力收割操作系统：注意力经济、营销、危机公关 | https://github.com/alchaincyf/sun-yuchen-perspective | - Network accessline 73
| zhang-yiming-skill | 张一鸣的认知操作系统：产品、组织、全球化、延迟满足 | https://github.com/alchaincyf/zhang-yiming-skill | - Network accessline 74
| taleb-skill | 塔勒布的认知操作系统：尾部风险、反脆弱、质疑主流叙事 | https://github.com/alchaincyf/taleb-skill | - Network accessline 75
| mrbeast-skill | MrBeast 的内容创造操作系统：视频标题/封面/钩子/留存 | https://github.com/alchaincyf/mrbeast-skill | - Network accessline 76
| paul-graham-skill | Paul Graham 的认知操作系统：创业、写作、做用户真正要的东西 | https://github.com/alchaincyf/paul-graham-skill | - Network accessline 77
| ilya-sutskever-skill | Ilya Sutskever 的认知操作系统：深度学习直觉、Scaling、AI 本质 | https://github.com/alchaincyf/ilya-sutskever-skill | - Network accessline 78
| huashu-art-motion | 艺术与解说动画：35种艺术风格、9种解说动画语法，拆解复刻、口播驱动整片，用代码让画动起来 | https://github.com/alchaincyf/huashu-art-motion | - Network accessline 79
| huashu-flash | 闪电.skill：照Anthropic两周把claude.ai提速3倍的方法给网站或App提速，先量基线、再证指标、逐步优化，用棘轮锁住只许变好 | https://github.com/alchaincyf/huashu-flash | - Network accessline 80
| mcstomach | 麦门胃：读麦当劳历史订单，带证据提炼口味规则，确认后生成可携带的个人「胃」 | https://github.com/alchaincyf/mcstomach | - Network accessline 86
- 公众号：「花叔」（在微信里搜「花叔」） - B站：花叔v：https://space.bilibili.com/14097567 - Network accessline 87
- B站：花叔v：https://space.bilibili.com/14097567 - Network accessline 88
- 小红书：花叔：https://www.xiaohongshu.com/user/profile/5abc6f17e8ac2b109179dfdf - Network accessline 89
- 个人网站：https://www.huasheng.ai/ - Network accessline 90
- X：https://x.com/AlchainHust - Network accessline 91
- YouTube：https://www.youtube.com/@Alchain - Network accessline 92
- 更多开源项目：https://github.com/alchaincyf?tab=repositories
references/生产流水线.md
- Network accessline 157
- `file://` 协议下同源本地数据的 `fetch()` 会失败，图表库常踩这个
references/视觉系统.md
- Network accessline 232
单文件 HTML 交付物额外注意：`file://` 协议下同源本地数据的 `fetch()` 会失败，数据要内联。
How the signals fit together
Facts from a static read of the files at the pinned commit. They describe what the code does, not why.
Declared and executed: 1 fact
- Contacts www.w3.org, which SKILL.md, the README and the plugin manifest do not mentionassets/chart.py:132High confidence
Signal confidence: 154 facts
- Network access in code of documentationREADME.md:24Low confidence
- Network access in code of documentationREADME.md:25Low confidence
- Network access in code of documentationREADME.md:27Low confidence
- Network access in code of documentationREADME.md:28Low confidence
- Network access in the frontmatter of the SKILL.mdSKILL.md:4Low confidence
- Network access in the frontmatter of the SKILL.mdSKILL.md:5Low confidence
- Network access in the frontmatter of the SKILL.mdSKILL.md:6Low confidence
- Network access in the frontmatter of the SKILL.mdSKILL.md:7Low confidence
- Network access in prose of the SKILL.mdSKILL.md:188Low confidence
- Network access in prose of the SKILL.mdSKILL.md:199Low confidence
- Executable script in code of a file that runsassets/chart.py:1High confidence
- Network access in code of a file that runsassets/chart.py:130High confidence
- Network access in code of a file that runsassets/chart.py:131High confidence
- Network access in code of a file that runsassets/chart.py:132High confidence
- Network access in code of a file that runsassets/chart.py:152High confidence
- Network access in code of a file that runsassets/chart.py:153High confidence
- Network access in code of a file that runsassets/chart.py:154High confidence
- Network access in code of a file that runsassets/chart.py:193High confidence
- Network access in code of a file that runsassets/chart.py:194High confidence
- Network access in code of a file that runsassets/chart.py:195High confidence
- Network access in code of a file that runsassets/chart.py:240High confidence
- Network access in code of a file that runsassets/chart.py:241High confidence
- Network access in code of a file that runsassets/chart.py:242High confidence
- Network access in code of a file that runsassets/chart.py:274High confidence
- Network access in code of a file that runsassets/chart.py:275High confidence
- Network access in code of a file that runsassets/chart.py:276High confidence
- Network access in code of a file that runsassets/chart.py:292High confidence
- Network access in code of a file that runsassets/chart.py:293High confidence
- Network access in code of a file that runsassets/chart.py:294High confidence
- Network access in code of a file that runsassets/chart.py:332High confidence
- Network access in code of a file that runsassets/chart.py:333High confidence
- Network access in code of a file that runsassets/chart.py:334High confidence
- Network access in code of a file that runsassets/chart.py:355High confidence
- Network access in code of a file that runsassets/chart.py:356High confidence
- Network access in code of a file that runsassets/chart.py:357High confidence
- Executable script in code of a file that runsassets/render.py:1High confidence
- Network access in prose of documentationcorpus/README.md:16Low confidence
- Network access in prose of documentationcorpus/README.md:17Low confidence
- Network access in prose of documentationcorpus/README.md:18Low confidence
- Network access in prose of documentationcorpus/samples.md:9Low confidence
- Network access in prose of documentationcorpus/samples.md:21Low confidence
- Network access in prose of documentationcorpus/samples.md:23Low confidence
- Network access in prose of documentationcorpus/samples.md:24Low confidence
- Network access in prose of documentationcorpus/samples.md:25Low confidence
- Network access in prose of documentationcorpus/samples.md:26Low confidence
- Network access in prose of documentationcorpus/samples.md:27Low confidence
- Network access in prose of documentationcorpus/samples.md:28Low confidence
- Network access in prose of documentationcorpus/samples.md:29Low confidence
- Network access in prose of documentationcorpus/samples.md:30Low confidence
- Network access in prose of documentationcorpus/samples.md:31Low confidence
and 104 more
Works with
| Agent | Status | Evidence | Install to |
|---|---|---|---|
| Claude Code | Compatible | Standard SKILL.md SKILL.md | .claude/skills/<skill-name>/ |
| Codex | Compatible | Standard SKILL.md SKILL.md | .agents/skills/<skill-name>/ |
| Cursor | Compatible | Standard SKILL.md SKILL.md | .cursor/skills/<skill-name>/ |
| Gemini CLI | Compatible | Standard SKILL.md SKILL.md | .gemini/skills/<skill-name>/ |
| GitHub Copilot | Compatible | Standard SKILL.md SKILL.md | .github/skills/<skill-name>/ |
| OpenCode | Compatible | Standard SKILL.md SKILL.md | .opencode/skills/<skill-name>/ |
- Native The repository ships a manifest, folder or frontmatter field for this agent.
- Compatible A standard SKILL.md that this agent can read.
- Unknown No evidence either way.
Derived automatically from paths, manifests and frontmatter. Not tested on each agent.
