Html
Skill y-a-v-a/skills/html
My agent skills
npx -y skills add y-a-v-a/skills --skill htmlAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Use this skill when the user wants to export the current "situation" — the state of the conversation, work done, findings, data, or context — to a standalone HTML file they can open in a browser. The agent decides what is most valuable to capture and visualize. Invoked as /html, optionally with hints about what to focus on.
The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.
SKILL.md
3.1 KB, as published. Nobody here has run it
HTML Export
Export the current situation to a self-contained index.html file in the
current working directory and give the user a clickable link to open it.
Optional focus hints
Arguments provided: $ARGUMENTS
If arguments are present, treat them as a hint about what to emphasize or
what kind of document to produce (e.g. /html focus on the architecture,
/html as a slide-style summary). If there are no arguments, you decide
everything.
Decide what to capture
This is the core of the skill: judge, from the current state of the conversation and context, what would be genuinely valuable to put in front of the user as an HTML page. You are not transcribing the chat — you are producing a useful artifact. Consider, and pick whatever fits:
- A summary of the work done, decisions made, and their rationale.
- Findings, analysis, or research results.
- Code changes, file structures, or architecture (diagrams welcome).
- Data worth visualizing as tables or charts.
- Status / checklist of what is done vs. outstanding.
- Anything else that is the real "payload" of this session.
Prefer signal over completeness. If one thing clearly dominates the
session, build the page around that. Give the page a meaningful <title>
and top-level heading that reflect the actual situation, not "HTML Export".
Build the file
- Produce a single, self-contained HTML5 document: all CSS inline in a
<style>block, no external stylesheets, fonts, scripts, or CDN links, so it opens correctly offline with a single click. - Make it clean and readable: a constrained content width (e.g.
max-width: 800px; margin: auto), comfortable line height, system font stack, sensible spacing, and styled tables/code blocks. Keep it tasteful, not flashy. - If a visualization genuinely helps (timeline, flow, simple chart), draw it with inline SVG or CSS rather than pulling in a library.
- Ensure the markup is valid and the document is complete
(
<!DOCTYPE html>,<html>,<head>with<meta charset="utf-8">and a viewport meta,<body>).
Choose the filename (never overwrite)
Write to index.html in the current working directory. If index.html
already exists, use index1.html; if that exists too, index2.html, and
so on — pick the first name in that sequence that does not yet exist. Never
overwrite an existing file.
Report the path
After writing the file, output its location as a file:// URL with the
absolute path, on its own line, so the user can cmd-click / ctrl-click it
to open it in a browser. For example:
file:///home/user/project/index1.html
Resolve the real absolute path of the working directory (do not hardcode the example above). Briefly mention what you chose to include and why.