Svg Audit logo

Svg Audit

Community
Sagargupta16
svg-audit

Audits an existing SVG file or animated React scene and returns ranked findings with fixes: render-target safety (GitHub README image proxy, browsers), accessibility, motion quality, file weight and code hygiene. Runs the bundled validator, then reviews what a linter cannot see. Use when the user says "audit this SVG", "why doesn't my SVG animate on GitHub", "check this README card", "is this SVG accessible", "review my animation", "shrink this SVG", or runs /svg-audit [file]. Reports first; only edits files when the user asks for the fixes.

Overview

PublisherSagargupta16
Repositorysvg-forge
Skill namesvg-audit
Stars
3
Forks
0
Bundled files
Instructions only
LicenseMIT
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 Sagargupta16 on GitHub. Read the source before you install it.

Installation

Install the Svg Audit 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.
    https://github.com/Sagargupta16/svg-forge/tree/main/skills/svg-audit
  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/Sagargupta16/svg-forge.git /tmp/svg-forge
mkdir -p .claude/skills
cp -r /tmp/svg-forge/skills/svg-audit .claude/skills/svg-audit
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Svg Audit 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 Svg Audit 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 Svg Audit 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.

svg-audit

Find what is wrong with an SVG or an animated scene, rank it by how much it hurts, and say exactly how to fix each item. Default mode is read-only: print the report and stop. Apply fixes only when the user asks ("fix it", "--fix", "apply").

1. Identify the artifact and its target

ArtifactTargetPrimary references
.svg shown through <img> in a README, docs site or emailstandalonegithub-readme, smil-and-css
.svg inlined in a web pagewebsmil-and-css, accessibility
React component animated with Motionscenemotion, verify

Ask only if the target cannot be inferred from where the file lives or how it is referenced (grep the repo for its filename).

The references above live in the sibling animated-svg and motion-scene skills of the svg-forge plugin; this skill reads them, so install the plugin, not this folder alone.

2. Run the validator (SVG files)

bash
python skills/animated-svg/scripts/validate_svg.py path/to/file.svg --target standalone --strict-contrast --json

Use --target web for inline page SVGs. For one half of a <picture> pair add --theme dark or --theme light; otherwise text with no background of its own is measured on both GitHub pages. Every error is a finding. Warnings become findings when they matter for the target. Quote rule ids (SV0xx) so the user can look them up; SV021-SV023 are WCAG contrast (text 4.5:1, large text and solid shapes 3:1), with each measurement under stats.contrast.

Then render the timeline so motion findings rest on frames, not guesses:

bash
python skills/animated-svg/scripts/frames.py path/to/file.svg --theme both

Read the PNG contact sheet it writes: entrances, overlaps, labels that are hidden at rest, and whether the seam frame matches t=0. Exit 3 means no browser was found; say so in the report instead of guessing.

For a scene component, there is no validator: read the component and its helpers instead, and apply the checks in section 3.

3. Review what the validator cannot see

Work through each lens and record findings with a file and line.

Render-target safety

  • Standalone: anything that needs network, script, or an HTML context fails silently on GitHub. External fonts fall back to system fonts, so check text still fits.
  • Dark and light: text needs 4.5:1 and marks 3:1 on both GitHub themes (the validator measures this), or the file must ship as a <picture> pair. In-SVG prefers-color-scheme does not follow GitHub's theme and fails on WebKit. For a fix, regenerate with forge.py --theme both or a preset; it tunes every color to pass.
  • Paint order: a far edge showing through a near panel, a wire crossing the face of the panel it should end at, a dashed guide over its own plate. The fix is order and an opaque fill, not a mask.
  • Browser quirks: SMIL path morphs are the only morphs Safari plays; calcMode misspelled (calcmod) silently drops keySplines; pathLength together with vector-effect="non-scaling-stroke" stops line draws short in Chrome.

Motion quality

  • Rest frame: read the static attributes as the first frame and judge it as the thumbnail. A flat grid, an empty tray or every element at opacity="0" fails; so does a rest that looks like nothing rendered yet.
  • One highlight: one bright mark at rest; a beat moves it, never adds a second. Two lit places at once is a finding.
  • Loop shape: a loop that opens on an empty stage leaves a half-built poster and frozen frame; opening on the finished state with a clear window, or a static ghost layer under the build-up, is the fix.
  • A state change: at least one beat decides something (a row rejected, a check popped, a slot filled); motion with nothing decided is a polish finding.
  • One clock: beats in a loop should share one duration so they never drift apart.
  • Seamless loop: the last keyframe equals the first, or the restart happens while the element is invisible.
  • Easing: multi-keyframe animations need per-segment easing (keySplines per segment, or an ease array in Motion when times is set).
  • Budget: count animated elements; more than about 12 in one scene, or 40 in one file, reads as noise and costs CPU.
  • Only transform and opacity (plus stroke dash properties for line draws) should animate in scenes.

Accessibility

  • Meaningful graphic: role="img" plus <title> (and <desc> when it carries information) referenced by aria-labelledby, or alt text on the <img>.
  • Decorative graphic: aria-hidden="true" (inline) or empty alt (image).
  • Auto-playing motion longer than 5 s has no pause control inside an image: note it, do not block on it.

Weight and hygiene

  • Size against the budget for the target (standalone: aim under 50 KB, hard stop 100 KB unless the user accepts more). python skills/animated-svg/scripts/optimize_svg.py FILE --check reports what an animation-safe minify would save without writing anything.
  • Embedded rasters at more than twice their drawn size, unused <defs>, duplicated paths that could be <use>.
  • Ids prefixed and unique when the file inlines icons or may be inlined into a page.
  • Deterministic output if the file is generated: same input, same bytes.

4. Report

Print a table ranked by severity, then a short verdict.

| # | severity | where | finding | fix |
| 1 | error    | card.svg:12 | <image href="https://..."> is blocked by GitHub's image proxy | inline it as a data URI under 64 KB, or drop it |
| 2 | warn     | card.svg:40 | 6 s loop with beats on 4 s and 6 s clocks drifts | put every beat on the 6 s clock |
verdict: renders on GitHub after fix 1; 2 is polish.

Severity: error (does not render or is unsafe), warn (renders but looks wrong somewhere or hurts access), polish (quality).

5. Fix mode (only when asked)

  • Apply fixes in severity order, one finding at a time, keeping unrelated markup byte-identical.
  • Re-run the validator after each change; the report's errors must reach zero.
  • For generated files, fix the generator or its input, not the output.
  • Show the diff and the new validator result.

Frequently asked questions

What does the Svg Audit AI skill do?

Audits an existing SVG file or animated React scene and returns ranked findings with fixes: render-target safety (GitHub README image proxy, browsers), accessibility, motion quality, file weight and code hygiene. Runs the bundled validator, then reviews what a linter cannot see. Use when the user says "audit this SVG", "why doesn't my SVG animate on GitHub", "check this README card", "is this SVG accessible", "review my animation", "shrink this SVG", or runs /svg-audit [file]. Reports first; only edits files when the user asks for the fixes.

Why use Svg Audit on TypingMind?

Because you install it once and use it with any model. Svg Audit 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 Svg Audit in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Sagargupta16/svg-forge/tree/main/skills/svg-audit. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Svg Audit?

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 Svg Audit?

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

Is the Svg Audit AI skill free?

Yes. It is published on GitHub by Sagargupta16 under the MIT 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 👇