masthead
Skill7 starsChanged 19 days agoMITView on GitHub
Build an academic project page for a research paper with the Masthead template (static HTML, dark mode, GitHub Pages ready). Use when the user wants a project page, paper website, or GitHub Pages site for a paper, or mentions Masthead. The skill reads the paper first, asks a few design questions, then builds and checks the page.
- Deletes or force-pushes
1 other project uses a version of this.
Adopted by
Most recently taken up by
- zanlib0/skillstheir own version, first seen Sep 13, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- zanlib0/skills★ 7same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
The file
We split this skill into the rules it gives an agent
10 rules. About 0 are common, 8 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈8
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 12
Ask the user, in one message, for anything not already given:
- Distinctive: line 19
If the user only gives an arXiv ID, fetch the abstract page and the source or PDF yourself.
- Distinctive: line 23
Read the whole paper before writing anything.
- Distinctive: line 38
Ask these in one round, each with your recommendation first.
- Very specific: line 42
Theme: default (Space Grotesk, blue), gazette (serif, oxblood), lab (IBM Plex, green) or mono (Geist, vermilion).
- Distinctive: line 45
Wait for the answers before building.
- Distinctive: line 63
Keep AGENTS.md in the new page.
- Very specific: line 73
Social preview: If the user agreed, serve the page, screenshot the hero at 1200 by 630 (headless Chrome works: --window-size=1200,630 …
- Distinctive: line 75
Never invent numbers, results, authors, venues or links.
- Distinctive: line 79
Run the checker that ships with this skill on the new page:
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Masthead: a project page for a paper
You turn a research paper into a finished project page built on the Masthead template (https://github.com/AntiQuality/masthead-project-page-template). Work in five phases and do not skip the questions in phase 3.
Phase 1: get the paper
Ask the user, in one message, for anything not already given:
- The paper: arXiv link, PDF or LaTeX source. LaTeX is best because figures and BibTeX come with it.
- The BibTeX entry, if it is not in the source.
- Links to show as buttons: code, data, video, slides, poster. Say which ones are not released yet.
- Where the page should live: a new GitHub repo (and its name), an existing repo, or a local folder.
If the user only gives an arXiv ID, fetch the abstract page and the source or PDF yourself.
Phase 2: read and analyze
Read the whole paper before writing anything. Collect:
- Title, authors in order, affiliations, equal contribution and corresponding author marks, venue and award (Oral, Spotlight, Best Paper).
- The one-sentence contribution and a one-line hook that contrasts the usual view with the paper's view.
- Three to six headline numbers for the stats strip, each with a short label. Only numbers stated in the paper.
- Three key findings, each a number plus one claim.
- The main results table, if there is one, and which columns are higher or lower is better.
- Data that would make a good bar chart (a few groups, two to four series).
- A natural before and after pair (comparison cards) and a qualitative example that compares methods (case study), if the paper has them.
- Figures: the teaser, the method overview and any result figures, with their captions.
Then show the user a short outline: the sections you plan, which Masthead module each one uses, and the numbers you picked. Mark anything you are unsure about.
Phase 3: ask the design questions
Ask these in one round, each with your recommendation first. Use your agent's question tool if it has one, otherwise a short numbered list.
- Title highlight. Which single word in the title gets the accent underline? Recommend the word that carries the idea.
- Modules. Which optional modules to keep: results table, bar chart, comparison cards, case study, video. Recommend only the ones the paper has real material for.
- Theme. default (Space Grotesk, blue),
gazette(serif, oxblood),lab(IBM Plex, green) ormono(Geist, vermilion). Recommend one that fits the field and the figures. - Logo and preview image. Does the project have a logo? Should you generate the 1200 by 630 social preview from the finished hero?
Wait for the answers before building.
Phase 4: build
Get the template. With the GitHub CLI and a new repo:
gh repo create <name> --template AntiQuality/masthead-project-page-template --public --clone rm -rf <name>/.github <name>/skillsOtherwise copy it into the target folder:
git clone --depth 1 https://github.com/AntiQuality/masthead-project-page-template <folder> rm -rf <folder>/.git <folder>/.github <folder>/skillsKeep
AGENTS.mdin the new page. It is the rulebook for every module.- Read
AGENTS.mdin the new folder and follow it for every edit. It covers the meta block, each module's markup, the text style and the constraints. - Fill
index.htmltop to bottom with what you collected in phase 2 and the answers from phase 3. Delete every module the user did not keep, renumber the section numbers and make the top bar links match the remaining section ids. - Figures. Copy them into
assets/media/. Convert PDF figures to PNG or SVG. Keep each image under about 500 KB, and set realwidthandheightattributes and descriptive alt text. Delete the placeholder files you replaced. - Theme. Link the chosen preset after
tokens.css, or leave the default. - Social preview. If the user agreed, serve the page, screenshot the hero at 1200 by 630 (headless Chrome works:
--window-size=1200,630 --screenshot) and save it asassets/media/social-preview.png. Pointog:imageandtwitter:imageat its absolute URL.
Never invent numbers, results, authors, venues or links. If something is missing, leave the module out or ask.
Phase 5: check and publish
Run the checker that ships with this skill on the new page:
python3 <this skill folder>/scripts/check_page.py <folder>/index.htmlFix everything it reports. It looks for leftover placeholders, forbidden punctuation, broken section links, missing alt text and social image problems.
- Serve the folder (
python3 -m http.server) and look at the page in light mode, dark mode and at phone width (about 375 px). Click the nav links, the resource buttons and the BibTeX copy button. - Show the user the result and list anything you left out and why.
Publishing changes things outside the user's machine, so ask before you push. With their go-ahead, commit, push and turn on GitHub Pages:
gh api -X POST repos/<owner>/<repo>/pages -f "source[branch]=main" -f "source[path]=/"The page goes live at
https://<owner>.github.io/<repo>/within a minute or two.
Text style
Masthead pages follow a strict style. In visible text, alt text, captions and comments: no em dashes, no en dashes, no middle dots and no semicolons. Rewrite with periods, commas, colons or "and". Write ranges as "1 to 4" and missing table values as n/a. The checker enforces this.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

