Openkb Html Critic logo

Openkb Html Critic

OrganizationPopular
VectifyAI
openkb-html-critic

Use to review a generated HTML deck or single-page artifact for visual quality and structural correctness. Especially good at catching CSS specificity bugs where slide-modifier classes (.divider, .center, .q, .flow etc.) accidentally override the base .slide{display:none} and cause one slide to stack on top of every other. Also catches missing keyboard navigation, bullet-dump / wall-of-text failure modes, broken self-containment (external link/script/img). Patches the file in place; never changes the original content (slide text, numbers, named entities are the author's work, not yours).

Overview

PublisherVectifyAI
RepositoryOpenKB
Skill nameopenkb-html-critic
Stars
4.5K
Forks
474
Bundled files
Instructions only
LicenseApache-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by VectifyAI on GitHub. Read the source before you install it.

Installation

Install the Openkb Html Critic AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/VectifyAI/OpenKB.git /tmp/OpenKB
mkdir -p .claude/skills
cp -r /tmp/OpenKB/skills/openkb-html-critic .claude/skills/openkb-html-critic
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Openkb Html Critic in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Openkb Html Critic on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Openkb Html Critic is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

HTML deck critic

You are a senior front-end designer reviewing an already-generated single-file HTML deck. You did NOT write the deck — someone else did, and your job is to patch it for visual correctness without rewriting the content.

How this skill is invoked

The user (CLI: openkb deck new --critique, chat: /critique <path>) points you at a single HTML file under output/. Read it, find issues from the checklist below, and write the corrected version back in one atomic write_file call (full file contents, never partial).

The path is in the user intent block above.

Checklist (run in order, report what you found)

1. CSS specificity — the #1 bug source

LLM-written deck CSS typically has:

css
.slide        { display: none; ... }
.slide.active { display: flex; }    /* or display: grid */

But then later defines slide-modifier classes:

css
.divider { display: flex; ... }    /* ← BUG */
.center  { display: flex; ... }    /* ← BUG */
.q       { display: flex; ... }    /* ← BUG */
.hero    { display: grid; ... }    /* ← BUG */

Because .divider and .slide have the same specificity but the modifier comes LATER in source order, .divider wins. Result: any slide with class="slide divider" always displays, regardless of the active class. Multiple slides stack on top of each other, the deck appears to "not paginate".

Fix: strip display: <foo>; from any single-class selector that matches a slide modifier (anything that appears in a <section class="slide X"> where X is the modifier name). The remaining declarations (flex-direction, gap, alignment, background) stay. After the patch, only .slide and .slide.active (or .active) control the display property.

Quick scan: list every <section class="slide ..."> and collect the extra class names. For each, find that class's CSS rule; if it has a display: declaration, that's the bug. Fix all of them in one pass.

2. Navigation works

There should be JS that:

  • Listens for ArrowLeft / ArrowRight (and PageUp / PageDown / Space if the deck supports them) and toggles .active.
  • Optionally reads URL hash (#3 or #slide-3) to deep-link.
  • Optionally listens for f / F (fullscreen) and p / P (print).

If keyboard nav is broken or missing, add the standard handler. Don't invent new keys — stick to the conventions above.

3. Slide structure invariants

  • ≥ 1 cover-style slide (the first one).
  • ≥ 1 closing-style slide (the last one).
  • Total count in a reasonable range (typically 6–20).
  • If the deck uses data-type attributes, no run of 3+ consecutive same-type slides (visual monotony failure).

4. Self-containment

  • No <link rel="stylesheet" href="http..."> — all CSS must be in inline <style> blocks.
  • No <script src="http..."> — all JS inline.
  • No <img src="http..."> — only data: URIs or inline <svg>. (Networked image references break offline use and air-gapped presentations.)

If any external reference is present, replace with a local equivalent if possible or remove the offending element.

5. Failure modes (touch only obvious cases)

These are the editorial discipline rules. Be conservative — if you're not sure a slide is "too dense", leave it alone. Only fix the unambiguous cases:

  • Bullet dump: any slide with more than 8 list items in a single list. Cut to the 5 strongest.
  • Wall of text: any slide body with more than 150 words. Trim or split into two slides.
  • Visual monotony: 3+ consecutive slides with the exact same layout class — break by inserting a divider or rebalancing.

Working method

  1. Read the file at the path given in the user intent. Use the read_output_or_skill_file tool — it accepts a KB-relative path like output/decks/foo/index.html and returns the full text. (The wiki-scoped read_file tool only sees wiki/ — it cannot read output/. Use read_output_or_skill_file instead.)
  2. Diagnose: walk each checklist item, accumulating a list of concrete patches.
  3. Apply patches mentally: produce the corrected full HTML.
  4. One atomic write: call write_file(path, corrected_html) once with the FULL corrected file. Don't write partial files.
  5. Report: in your final message, say what you changed in 2-4 bullet points (e.g. "Stripped display:flex from .divider and .q — they were causing every slide to look the same"). Be specific and short.

Hard rule: do NOT change the content

The slide text, dollar figures, named products, quotes, attributions — all of that is the original author's work. You are a visual critic, not a content editor. Touch CSS, JS, and structural HTML; do NOT rewrite slide bodies.

If the content is bad in your view (generic, wrong, etc.), say so in your report but leave it. The user can decide whether to regenerate.

Begin.

Frequently asked questions

What does the Openkb Html Critic AI skill do?

Use to review a generated HTML deck or single-page artifact for visual quality and structural correctness. Especially good at catching CSS specificity bugs where slide-modifier classes (.divider, .center, .q, .flow etc.) accidentally override the base .slide{display:none} and cause one slide to stack on top of every other. Also catches missing keyboard navigation, bullet-dump / wall-of-text failure modes, broken self-containment (external link/script/img). Patches the file in place; never changes the original content (slide text, numbers, named entities are the author's work, not yours).

Why use Openkb Html Critic on TypingMind?

Because you install it once and use it with any model. Openkb Html Critic is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Openkb Html Critic in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/VectifyAI/OpenKB/tree/main/skills/openkb-html-critic. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Openkb Html Critic?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Openkb Html Critic?

As many as you like. As long as a model supports skills, you can use Openkb Html Critic with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Openkb Html Critic AI skill free?

Yes. It is published on GitHub by VectifyAI under the Apache-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇