agentleFS
RuneStone-TechRuneStone-Tech/optimize-anything/skills/router/SKILL.md

optimizer-router

Skill1 starsChanged 8 days agoMITView on GitHub

Route an optimization request to the right skill in this repo when the user asks which optimizer or skill to use, is unsure where to start, names several technologies, or describes a symptom without naming a stack (slow app, memory leak, huge bundle, flaky deploy, architecture review). Use this to pick one primary skill plus at most one library/style/infra skill, then hand off to that skill's workflow. Do not use it when the user already named a specific stack or skill.

The file

We split this skill into the rules it gives an agent

9 rules. About 0 are common, 5 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If the user named a specific technology, skip the router and use that skill.

    line 20
  2. Distinctive:

    If the catalog is missing:

    line 30
  3. Very specific:

    Read skills/CATALOG.md for the full grouped list.

    line 36
  4. Very specific:

    Architecture: only if the problem is boundaries/data/distribution, not a local performance bug.

    line 64
  5. Distinctive:

    Do not run three audits blindly.

    line 75
  6. Distinctive:

    If the main stack is still ambiguous, ask exactly one question:

    line 79
  7. Very specific:

    Do not ask about every library.

    line 83
  8. Very specific:

    Read the chosen skill's SKILL.md and follow its workflow.

    line 87
  9. Distinctive:

    Do not summarize a skill's advice from memory.

    line 93

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Optimizer Router

The repo has 64+ focused skills. Do not guess from memory and do not read the whole tree. Use the catalog, pick the smallest set, then follow the target skill.

1. Find the signals

From the request and the repo, detect:

  • Framework/runtime: React, Next.js, Vue, Nuxt, Angular, Svelte, Solid, Astro, Remix, TanStack Start, Node, or a backend language.
  • Library/style: component kit (MUI, Mantine, shadcn, ...), styling system (Tailwind, CSS, Sass, CSS-in-JS), motion/3D.
  • Infra: Docker, Kubernetes, Terraform.
  • System shape: monolith/modular/microservices/serverless, shared DB, coupling.
  • Symptom: slow load, slow interaction, memory, bundle size, database, reliability, cost, migration.

If the user named a specific technology, skip the router and use that skill. The router is only for ambiguity.

2. Search the catalog

python tools/catalog.py search "slow nextjs mui table"
python tools/catalog.py search "postgres connection pool" --limit 5
python tools/catalog.py list --domain backend

If the catalog is missing:

python tools/catalog.py

Read skills/CATALOG.md for the full grouped list.

3. Route by domain

SignalPrimary skill
React / Next.js / Remix / TanStack Startskills/web/react, nextjs, remix, tanstack-start
Vue / Nuxtskills/web/vue, skills/web/nuxt
Angular / Svelte / Solid / Astroskills/web/angular, svelte, solid, astro
Node HTTP/API/CLIskills/web/node
Node backend service (SLOs, shutdown, K8s)skills/backend/node
Rust / Go / C / C++ / C# / Rubyskills/backend/<language>
Component libraryskills/web/_components/skills/<library>
Styling systemskills/styles/<system>
Motion / 3D / animationskills/motion/<library>
Docker / Kubernetes / Terraformskills/devops/<tool>
Coupling, boundaries, shared DB, monolith vs servicesskills/architecture
Generic UI library comparison; no focused skillskills/web/_components

Focused libraries live under: skills/web/_components/skills/ (Aceternity, Base UI, Bulma, coss, Create UI, DaisyUI, Float UI, Flowbite, Franken UI, Headless UI, HeroUI, Mantine, MUI, Park UI, Preline, Radix, ReUI, shadcn, TailGrids, tweakcn, Untitled UI).

Styling systems: bootstrap, pandacss, sass-scss, stylex, tailwindcss, twind, unocss, vanilla-extract.

4. Pick at most 3

  • Primary: the framework/runtime/language.
  • Secondary: the component library, styling system, motion/3D or infra that is actually involved.
  • Architecture: only if the problem is boundaries/data/distribution, not a local performance bug.

Examples:

  • “Slow Next.js product page with MUI tables” → skills/web/nextjs + skills/web/_components/skills/mui.
  • “Vue app with Tailwind feels sluggish” → skills/web/vue + skills/styles/tailwindcss.
  • “Rails endpoint N+1” → skills/backend/ruby.
  • “Docker image is 2 GB” → skills/devops/docker.
  • “Should we split this monolith?” → skills/architecture (+ framework skill only if implementation detail matters).
  • “Billing service shares the orders DB” → skills/architecture + skills/backend/node or the relevant language.

Do not run three audits blindly. Pick the primary constraint; add a second skill only when its layer is part of the problem.

5. Ask one question only if needed

If the main stack is still ambiguous, ask exactly one question:

What is the main app stack: React/Next.js, Vue/Nuxt, Angular, Svelte, Solid, Astro, Remix, TanStack Start, a backend language, or infrastructure?

Do not ask about every library; inspect package.json, manifests, lockfiles and the tree first.

6. Hand off

Read the chosen skill's SKILL.md and follow its workflow. Tell the user:

  1. which skill(s) you chose,
  2. why (one line),
  3. what you will run first (the skill's audit command).

Do not summarize a skill's advice from memory; open the file.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.