huashu-report skill
A skill that runs an agent through researcher, editor and designer roles to produce a report in the style of top institutions.
- 493 ★ on the repository
- Last commit
- MIT
Minimal access
Why this level
- The skill is text instructions and three local Python scripts, with no external services or secrets.
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/huashu-reportDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewMinimal accessThe tool runs nothing and makes no network calls.
- The skill is text instructions and three local Python scripts, with no external services or secrets.
This is third-party code. Review the repository files before installing.
What it does
huashu-report is a skill for institution-grade research reports: industry reviews, white papers, annual surveys, data insights and arXiv papers. The conventions were derived from analysing 41 reports by Stanford HAI, McKinsey, BCG, OpenAI, Anthropic, PwC, Deloitte and others. The agent works in turn as researcher, editor, information designer and data visualiser. `SKILL.md` asks three questions up front and picks one of six report types. Reference files cover structure, sentence style, research depth, the visual system and 8 chart types, while `assets/chart.py` and `assets/render.py` draw SVG charts and assemble a report with automatic checks. No external services are needed.
Who it is for. Analysts, researchers and white paper authors who delegate reports to an agent and want structure and charts at major-institution level.
Good fit when
- An industry report or white paper needs the structure used by major consulting firms.
- An arXiv paper is formatted in IMRaD and LaTeX.
- Data becomes charts with conclusion titles and a source line.
- A report for a lay reader follows a three-act skeleton.
Not a fit when
- You need a single article or an ordinary slide deck: the README treats those as other work.
- The norms come from English reports, so Russian and Chinese writing habits need a manual check.
- The topic is far from AI and technology: the sample leans that way and other fields differ.
Example request
Prepare a white paper on the AI agent market with three charts, source lines and a limitations section.Limitations
The skill is written in Chinese. The 41-report sample leans toward AI and technology, and other fields may differ. The norms come from English reports, so Russian text needs its own proofreading. Assembly needs Python; the corpus reports themselves are not in the repository, only the source list and analysis figures. The agent runs three Python scripts, so read them before the first run.
How to disable. Delete the `huashu-report` folder from the agent's skills directory (`~/.claude/skills/` or `~/.agents/skills/`).
FAQ
How do I install it for several agents at once?
Clone into a shared pool: `git clone https://github.com/alchaincyf/huashu-report ~/.agents/skills/huashu-report`.
How does the skill start?
No explicit call is needed: a request such as "make a research report" or "prepare a white paper" triggers it.
What is in the assets folder?
`chart.py` draws 8 SVG chart types with negative values, `render.py` assembles the report, fills page numbers into the contents and checks the result, and `base.css` sets styles.
Where do the norms come from?
From 41 analysed reports: Stanford HAI, McKinsey, BCG, OpenAI, Anthropic, PwC, Deloitte, World Bank and Reuters Institute. The source list and figures are in `corpus/`.
Is it suitable for slide decks?
No. The README places single articles and decks outside its scope. The exception is a 16:9 consulting-style report described in `SKILL.md`.
Discussion
No questions or reviews yet.
Related
SkillsA skill turns an ML research repository into a conference paper draft in LaTeX with a claim-evidence map.
A document pipeline skill: anything to markdown, markdown to clean html and publisher-grade docx, and html back to markdown
Skill with 30 prompt templates for research papers in English and Chinese, from literature review to grant proposals
The agent turns rich answers, files, folders and service exports into an HTML page in a ready style, checked in a browser
A library of 280+ research skills: literature search, experiment design, analysis, reference checks and paper preparation
16 ready skills for Cursor, Claude Code and Codex: theses, PDF and Office, draw.io diagrams, project documentation
README in short
The skill is plain-text references plus three Python files, with no service dependency, and works across agents. Install it by cloning into a skills directory.