Transitions Agent logo

Transitions Agent

CommunityPopular
Jakubantalik
transitions-agent

Scan a codebase for missing, janky, or inconsistent UI motion and fix it with transitions.dev recipes. ALWAYS use this skill when the user's message is or contains a transitions-agent command (npx transitions-agent, npx transitions-agent fix, any variant) - the command is a request for the full guided flow, not just command execution. Also use when the user asks about UI transitions, animations, motion quality, a motion score, janky or missing transitions, "polish the motion", "fix the animations", or whether the motion is good after UI changes.

Overview

PublisherJakubantalik
Repositorytransitions.dev
Skill nametransitions-agent
Stars
5.9K
Forks
279
Bundled files
Instructions only
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 Jakubantalik on GitHub. Read the source before you install it.

Installation

Install the Transitions Agent 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/Jakubantalik/transitions.dev/tree/main/agent/skill
  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/Jakubantalik/transitions.dev.git /tmp/transitions.dev
mkdir -p .claude/skills
cp -r /tmp/transitions.dev/agent/skill .claude/skills/transitions-agent
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Transitions Agent 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 Transitions Agent 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 Transitions Agent 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.

Transitions Motion Agent

You drive the transitions-agent CLI for the user: scan, present choices, run the choice they pick, verify. The scan is deterministic and free; fixes change files, so every fix step needs the user's go-ahead first.

When the user's message is just the scan or fix command (npx transitions-agent, npx transitions-agent fix, or a variant), that IS the request for this whole workflow: run the scan, then ALWAYS continue to the options step below - never stop at reporting the scan output. init-ci is different: it is CI setup only (see CI setup) - never scan or fix as part of it.

Workflow

  1. Scan (safe, read-only, no account needed) from the repository root:

    bash
    npx transitions-agent@latest --json

    Report to the user: the motion score (0-100), then the components list - every UI component the scan recognized (modal, dropdown, tooltip, toast, accordion, tabs, toggle...) with its motion and the transitions.dev recipe it maps to, marked as matching the recipe, off the motion scale, built wrong, or hand-rolled - then the findings grouped by rule with file:line. Keep it short; lead with the score.

  2. Present exactly these two options and let the user choose (do not fix without a choice, and do not invent other options such as fixing the findings yourself):

    • Polish fix - the transitions-polish treatment: every value moved onto the motion-token scale by what it does (a 300ms modal close becomes 150ms), hover transitions covering what the hover changes, a reduced-motion guard. Small diffs, nothing restructured: npx transitions-agent@latest fix --yes
    • Revamp fix - everything polish does, plus the transitions.dev recipe installed on each recognized component (Pro recipes included): recipe CSS, state hooks, and the JS for enter and exit. Business plan: npx transitions-agent@latest fix --mode revamp --yes
  3. License: hosted fixes need a key. It resolves from TRANSITIONS_AGENT_LICENSE, or ~/.transitions-agent.json. If neither exists, offer:

    bash
    npx transitions-agent@latest signup

    This opens the user's browser; they enter their email there and the key saves itself. Wait for it to finish, then run the fix. If it reports the key could not be saved (sandboxed agent), pass it on every fix with --license <key>.

  4. Verify: re-run the scan and report the score change (for example "71 to 89"). If the user wants it as a pull request, use fix --pr (it creates a branch and opens a PR; it never touches the main branch) - only with their explicit ok.

CI setup

When the user asks to add the Agent to GitHub Actions / CI, or their message is npx transitions-agent init-ci (any flags), set up CI and nothing else: do NOT run a scan or a fix. Run:

bash
npx transitions-agent@latest init-ci

It writes both workflows (score plus automatic fixes on every pull request: a fix pull request into that branch, never main, that the user merges to apply or closes to reject; and a manual fix run from the Actions tab), keeps existing ones, checks the GitHub side, shows which fix modes the user's key allows, and ends with lettered Options. Relay a one-line status, then present those options exactly as printed and ask which to do. Automatic fixes default to polish; ask the user which they want when it matters: --mode revamp (Business) or --no-auto-fix (scores only). Labels on a pull request override it per pull request: revamp, polish, no-motion-fix. Typical options:

  • Sign up or sign in (no key yet): npx transitions-agent@latest signup opens the browser; existing accounts sign in the same way. Wait for it, then re-run init-ci.
  • Finish the GitHub side (license secret + pull request permission): npx transitions-agent@latest init-ci --yes. The key never appears in output.
  • Revamp as the default fix mode (Business keys): npx transitions-agent@latest init-ci --mode revamp.
  • Turn on automatic fixes for an older workflow: npx transitions-agent@latest init-ci --upgrade.
  • Commit the workflow files: on a new branch with a pull request, never straight to main.
  • Create a GitHub repository when the project has none.

Run only what the user picks. If gh is missing or blocked by a sandbox, give the user the manual steps it printed. Finish when it says "CI is fully set up", or tell the user exactly which step is theirs.

Network

transitions-agent needs network access: npm has to fetch the current version, and fix, signup, and init-ci --yes talk to api.transitions.dev and GitHub. A sandbox with network off (Codex runs commands that way by default) shows up as one of these:

  • npm error code ETARGET / "No matching version found" for a version that exists on npm,
  • a report header or --json version older than the latest release (npm served a stale cached copy),
  • "Could not reach the fix service", ENOTFOUND, or EAI_AGAIN.

That is the sandbox, never a missing package. Re-run the same command with network access (request escalated permissions or network approval from the user) and tell the user transitions-agent needs network. Never report the package as unavailable, and never present results from an older cached version. Mention the version from the report header when you report a scan.

Rules

  • Never run fix without the user choosing to, and never apply motion fixes yourself instead of the service - fixing is the product.
  • Never push to the main branch; --pr flows are review-first by design.
  • Free plan is polish-only; if revamp is refused by the service, relay the upgrade path (transitions.dev/pro.html) without pushing.
  • Motion-only changes: if a proposed fix would touch component logic, stop and tell the user.

Frequently asked questions

What does the Transitions Agent AI skill do?

Scan a codebase for missing, janky, or inconsistent UI motion and fix it with transitions.dev recipes. ALWAYS use this skill when the user's message is or contains a transitions-agent command (npx transitions-agent, npx transitions-agent fix, any variant) - the command is a request for the full guided flow, not just command execution. Also use when the user asks about UI transitions, animations, motion quality, a motion score, janky or missing transitions, "polish the motion", "fix the animations", or whether the motion is good after UI changes.

Why use Transitions Agent on TypingMind?

Because you install it once and use it with any model. Transitions Agent 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 Transitions Agent in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/Jakubantalik/transitions.dev/tree/main/agent/skill. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use Transitions Agent?

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 Transitions Agent?

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

Is the Transitions Agent AI skill free?

It is published on GitHub by Jakubantalik. Check the repository for licensing terms. 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 👇