More screenshots

    Works with the AI Agents you already use

    Claude CodeClaude CodeCursorCursorCodex CLICodex CLIDeepSeekDeepSeekMuseMuse+22 more

    HTML Explainer Pages: reports and plans as one file

    1

    Your agent writes reports, plans and reviews as one polished HTML page, then lints and screenshots it.

    $12

    Secure checkout via Stripe

    Money-back guarantee

    Converts to your local currency at checkout

    Security scanned
    HTML Explainer Pages: reports and plans as one file

    HTML Explainer Pages: reports and plans as one file

    Example session with this skill installed

    Write up this PR review as a one-page HTML explainer with the findings ranked by severity

    • Read your context and instructions
    • Compiled the html explainer pages
    • Generated the app files

    The review is now one self-contained HTML file, review.html, built from the light starter. It opens with the verdict and ranks all six findings by severity.

    How the page is laid out

    Dock    Findings · Offsets · Week 53 · Others · Verdict
    Hero    "Request changes. Two small bugs block the merge."
    01      Findings table, ranked, with status chips
    02      Must-fix: negative offsets lose the sign on the minutes (diff + callout)
    03      Must-fix: week 53 accepted in 52-week years (diff + callout)
    04      The four smaller items as open rows
    05      Night section: the verdict, what must land, and a PR comment with a Copy button
    

    The findings, as ranked on the page

    #SeverityFindingWhere
    01Must-fix-03:30 parses as -02:30offsets.py:41
    02Must-fixWeek 53 accepted in 52-week yearsiso_week.py
    03Should-fixNo year-boundary teststests/
    04Should-fixRegex compiled on every callparse.py
    05Should-fixREADME shows the old signatureREADME
    06OptionalError message says only "invalid date"parse.py

    What I ran

    The inliner's tests passed, 16 of 16, and all 14 kit files matched the hashes in SKILL.md. The inliner reported inlined review.html [FONTSx1, LOGOx2, KITx1, JSx1] and a second run said unchanged. The finished file is 91,230 bytes with the kit, script and logo inside it.

    layout_lint.py reported 0 issues at 1280px and 0 at 390px. qa_shots.py clicked every dock link and each one found its section, confirmed the fonts loaded at both widths, and saved four screenshots, which I looked at one by one.

    Next steps

    • Paste the comment from section 05 into the PR.
    • If you want a PDF to attach, run python3 vellum/scripts/html_to_pdf.py review.html.

    review.html

    ZIP · project files

    Generated

    Example file from a real run - the skill writes it into your workspace.

    Connects securely to your tools. The creator never sees your data.

    About this skill

    An agent can write a long report. Making one people want to read is harder. Markdown loses the reader after a screen, and a page styled from scratch looks generated and breaks on a phone.

    This skill gives your agent a finished design kit and a way to check its own work. It writes a single self-contained HTML file: a CSS kit and a small script inlined into the page, with widgets built for explaining things. There is a decision path for procedures, a before and after split, ledger arithmetic with a source line under every number, metric gauges against a target, a timeline, quote cards and tables. It also draws SVG concept figures by hand, no chart library. Pages come in light or dark, print cleanly to PDF, and take your accent colour and logo from a small theme file.

    Before the page is handed over, a lint runs at desktop and phone widths and flags text too close to an edge, overflow, text under 13px, overlapping labels, weak contrast and dock links that go nowhere. A screenshot pass then clicks the navigation, reports console errors and font loading, and saves images for the agent to look at. An optional PDF step catches the grey-box artefacts that blurred shadows leave in print, and pages left nearly blank.

    You get SKILL.md, the kit (CSS and JS), light and dark starters, a worked example page, markup references, a guide to drawing figures, and five Python scripts with 16 unit tests for the inliner.

    It needs Python 3.9 or later. The lint, screenshots and PDF need Playwright with Chromium, and the PDF checks use PyMuPDF if it is installed. Fonts load from Google Fonts. It suits anyone whose agent writes audits, plans, research summaries, code reviews or handover docs for other people.

    Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Technical Documentation.

    It builds static pages, not web apps. The lint, screenshots and PDF need Playwright with Chromium. Fonts load from Google Fonts, so a page opened offline falls back to Helvetica or Arial. The colour tints need a browser from 2023 or later (CSS color-mix).

    How to use

    1. 1

      Connect your agent to the Agensi MCP

      Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup

    2. 2

      Sign in to Agensi

      Your agent shows a sign-in link. Sign in and click Allow.

    3. 3

      Ask for what you need

      Your agent loads this skill and uses it on your task. It always asks before buying a skill.

    Reviews

    No reviews yet

    Be one of the first to try it. Every listed skill goes through the trust checks below.

    Security scanned

    Checked by our 8-point scan before listing

    Fresh listing

    Recently published to Agensi

    Money-back guarantee

    Doesn't do what its page promises? You get your money back.

    Trust & safety

    Security scan

    Security scan passed 13 days ago

    Software
    Python 3.9+PlaywrightChromium
    Optional
    PyMuPDFPillow

    Playwright with Chromium runs the lint, screenshots and PDF. PyMuPDF and Pillow are for the PDF checks. Pages load Google Fonts when online.

    Permissions

    Terminal / Shell
    Read Files
    Write Files
    Browser
    Listed13 days ago

    What's inside

    Frequently Asked Questions

    Python 3.9 or newer runs the inliner and its tests with the standard library alone. The layout lint, screenshots and PDF need Playwright with Chromium. PyMuPDF and Pillow are optional and add the PDF checks; without them the PDF is still written.

    None of the scripts calls a web service. They read your page and write HTML, PNG and PDF files on your disk. The page itself links to Google Fonts, so the browser fetches the two typefaces from Google whenever the page is opened online, including during the lint and screenshots. Offline it falls back to Helvetica or Arial.

    The agent never writes the CSS. A script copies in a fixed kit, so every page gets the same tested styles. Then a lint measures what a model cannot see in its own markup: text under 13px, contrast under 4.5:1, overflow at phone width, and dock links pointing at sections that do not exist.

    The inliner ships with 16 unit tests, and all 16 passed in our demo run. The demo page then linted clean at 1280px and 390px, and the screenshot pass found every dock link's section. During the build, the worked example's PDF showed no grey-box artefacts or near-blank pages.

    Yes. A small JSON theme file sets the accent colour and points at your SVG logo, and the tints follow from the accent. The inliner warns when the accent falls under 4.5:1 contrast and refuses a logo that contains a script. A misspelled token name stops the run, so a typo cannot slip through.

    The kit uses CSS color-mix(), which needs Chrome 111, Safari 16.2, Firefox 113 or later. Older browsers lose the accent tints. The page carries no chart library and no icon set: figures are drawn in SVG by hand, and the only outside request is for the fonts.

    New in Technical Documentation