feat(docs): add SKILL.md for agent integration of scaffold-ui#105
Merged
Conversation
Captures the integration gotchas the docs don't shout about loud enough: peer-dep ordering between @scaffold-ui/components and @scaffold-ui/hooks, the styles.css import as a silent failure mode, provider wrapping order, ssr:true for Next App Router, the chain defaulting rule, style prop memoization, and when to opt into @scaffold-ui/debug-contracts. Auto-picked up by gen-agent-skills-index.ts at build time and surfaced via /.well-known/agent-skills/index.json (RFC v0.2.0).
Adds /SKILL.md to the Link response header so agents discover it on the same HEAD request that surfaces llms.txt and the agent-skills index.
|
The latest updates on your projects. Learn more about Vercel for GitHub.
1 Skipped Deployment
|
technophile-04
commented
May 6, 2026
rin-st
approved these changes
May 6, 2026
Member
rin-st
left a comment
There was a problem hiding this comment.
Working great!
Test plan checkpoints: ✅
Vibe test:
Asked to add AddressInput and EtherInput to the main page of the pure next.js app. Everything was installed right, but EtherInput wasn't working first because default http() transports with eth.merkle.io was failing. So EtherInput was disabled. I asked claude just like "EtherInput is disabled, fix it" and then it found the bug and fixed it. After that I asked to add debug page, and it added the page for mainnet WETH. Not sure why WETH but it works as expected
technophile-04
commented
May 7, 2026
Contributor
Author
technophile-04
left a comment
There was a problem hiding this comment.
Merging this! thanks @rin-st for the review
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Lands the
SKILL.mdthat the discoverability infra in #104 was already designed to pick up. With this in,https://ui.scaffoldeth.io/SKILL.mdbecomes the agent-facing integration guide and/.well-known/agent-skills/index.jsongoes from{ "skills": [] }to a single-entry index pointing at it.The SKILL.md is deliberately scoped to what the docs don't shout about loud enough — not a re-documentation of every prop and hook (those already live in
llms-full.txt). It captures the integration gotchas that bite agents in practice:@scaffold-ui/componentsand@scaffold-ui/hooksstyles.cssimport as a silent failure modeWagmiProvider→QueryClientProvider→ optionalRainbowKitProvider)ssr: truefor Next.js App Routerchain={mainnet}everywherestyleprop memoization@scaffold-ui/debug-contractsand when to skip itThe first line of the body redirects agents to
llms-full.txtfor full API docs, so the SKILL stays focused and small (~110 lines body, well under Anthropic's 500-line guidance).Description field
Optimised for triggering accuracy following Anthropic's skill-authoring best practices:
Files changed
docs/public/SKILL.md(new, 117 lines including frontmatter)vercel.json(+1/-1) — adds</SKILL.md>; rel="alternate"; type="text/markdown"to the Link header so agents discover it on the same HEAD request asllms.txtand the skills indexTest plan
pnpm buildfromdocs/— confirmdocs/dist/SKILL.mdexists anddocs/dist/.well-known/agent-skills/index.jsonshows 1 skill entry with a valid sha256 digestcurl -s https://ui.scaffoldeth.io/SKILL.md | head -5returns the frontmattercurl -s https://ui.scaffoldeth.io/.well-known/agent-skills/index.json | jq '.skills | length'returns1curl -sI https://ui.scaffoldeth.io/ | grep -i ^linkshows the new SKILL.md entry alongside the existing fourshasum -a 256 <(curl -s https://ui.scaffoldeth.io/SKILL.md)should match the digest in the indexWhy this is a separate PR from #104
Kept the discoverability infra purely structural so reviewers could verify the scripts and headers in isolation. This PR is the content that flows through that infra. They could have shipped together but stacking made the diffs cleaner to review.
Out of scope (future)
/SKILL.md); could add/build-with-ai/skilllater if humans want to read it nicely in the docs UI