Hardcoded post -> md post rendered in jsx

Author: Pabl0cks

packages/nextjs/content/blog/ai-native-ethereum-stack.md

@@ -0,0 +1,81 @@
+---
+title: "Building an AI-Native Ethereum Developer Stack"
+date: "February 2026"
+description: "How we're restructuring our Ethereum developer stack: Scaffold-ETH 2 and Speedrun Ethereum to be AI native."
+---
+
+For years, we've been building our dev tools and learning materials primarily for humans. Docs explained what our toolkit did. We taught developers with guided, hands-on experiences. Extensions shipped as mergeable code.
+
+But we're finding that this model gets a bit clunky when an AI agent is driving the keyboard.
+
+Instead of just asking "how do we add AI to our tools" we started asking ourselves: what would our stack look like if we treated the AI as a primary user alongside the developer?
+
+For us, that meant rethinking some of our core primitives rather than just patching on new features. We've been doing exactly that across our dev toolkit ([Scaffold-ETH 2](https://github.com/scaffold-eth/scaffold-eth-2)) and our Solidity curriculum ([Speedrun Ethereum](https://speedrunethereum.com)). The Sand Garden, an intentional team within BuidlGuidl, is executing this vision.
+
+Here is a look at how we're restructuring our own Web3 stack for this shift.
+
+## Structuring Docs for Agents
+
+We removed Cursor rules from Scaffold-ETH 2 and replaced them with `AGENTS.md`.
+
+Cursor rules only work in Cursor. `AGENTS.md` is picked up by Claude Code, Cursor, Windsurf, and any other agent harness. **One file, every tool.** Every conversation starts with the full stack context already loaded.
+
+In these new workflows, we're noticing the main consumer of our docs is often an agent loading context before writing code. So we shipped `llms-full.txt`: the entire SE-2 documentation as a single flat file. Not a website. A file an agent loads into context and reasons against.
+
+It's the exact same information, just formatted for how an AI actually consumes it.
+
+## Swapping Complex Code for AI Skills
+
+Adding a Scaffold-ETH extension used to mean resolving `package.json` conflicts through hundreds of lines of template processing code.
+
+We replaced it with `/add-extension`: a simple agent skill built to work across different harnesses. Node.js handles the deterministic operations (fetching, copying), while the AI handles the judgment calls (merging). **Hundreds of lines of template code became a markdown file and a focused script.**
+
+Same pattern for developer workflow. Our `pr-create` skill is a markdown file that tells an agent how to inspect the diff, format the PR body, and open it via `gh`. No custom script. No alias. Just context.
+
+We've started stripping out custom scripts wherever we find that a model just naturally handles the task better.
+
+## Rethinking How We Teach
+
+Speedrun Ethereum is how many developers learn to build on Ethereum. We're rebuilding the learning layer to the same standard.
+
+**Per-challenge context files.** Each challenge gets an `AGENTS.md` detailing the challenge overview, smart contract structure, and frontend architecture. When a learner opens a challenge in their IDE, the agent already knows the environment.
+
+**AI Teacher Mode.** Type `/start` in the directory and an agent walks you through it. It asks questions, checks your understanding, guides you without giving the answer away, and reviews your code at your pace.
+
+Build ideas are also migrating. Old format: a GitHub repo link. New format: a prompt you paste into your agent. The agent reads context and starts building with you.
+
+## The Research Layer
+
+The agent ecosystem shifts weekly. That's why we aren't assembling heavy wrappers. We need to understand the raw primitives so we can adapt the moment a new paradigm drops.
+
+[raked](https://github.com/BuidlGuidl/raked) is a minimal TypeScript agent built for exactly this: the agent loop, sessions, memory, tools, skills. It's under 100 lines for the core. Built to be read, understood, and extended to solve your specific use cases.
+
+Alongside it: an experimental RAG pipeline on Arbitrum DAO governance data. Vector search via `pgvector`, retrieval via LlamaIndex, automated evaluation scoring Faithfulness, Relevancy, and Correctness.
+
+We built both because we really wanted to understand the underlying mechanisms before building heavier tools on top of them.
+
+## Figuring Out the Boundaries
+
+Security stays deterministic. Wallet interactions, transaction signing, and key management require hard boundaries. An agent making onchain transactions autonomously is an attack surface. That problem isn't solved yet.
+
+But **code review** is a perfect fit for AI judgment.
+
+Take [Grumpy Carlos](https://github.com/technophile-04/grumpy-carlos-personality-fetcher): a Claude Code subagent with a review personality inferred from scraped BuidlGuidl PR history. Drop it in `.claude/agents/`, ask for a review, and it responds exactly the way Carlos would: specific, strict, no vague feedback.
+
+The right context, structured for an AI to use, produces better output than general AI applied to problems without context.
+
+## The Compounding Effect
+
+To us, this is what going "AI-first" actually looks like in practice. It's not just dropping a Copilot plugin into Scaffold-ETH. It's a stack where every layer—the framework, the docs, the extensions, the curriculum—is structured to be used by an agent, not just tolerated by one.
+
+An agent building with SE-2 understands the stack. An agent teaching on Speedrun Ethereum understands the challenge. With `/add-extension` the agent can install new capabilities without leaving the conversation.
+
+**Next up:** an open-source plugin that packages all of this into one installable toolkit for Claude Code, OpenCode, and Cursor.
+
+## Try It Today
+
+The PRs are open and the code is live. If you're building at the intersection of AI and Ethereum:
+
+- **Start with raked:** [Read the agent loop.](https://github.com/BuidlGuidl/raked) Understand what a tool call actually is before building on top of abstractions.
+- **Look at Scaffold-ETH 2:** [See the AGENTS.md and /add-extension skills in progress.](https://github.com/scaffold-eth/scaffold-eth-2) Build your first Dapp with it and your favourite agent.
+- **Try the AI challenges:** Available on [Speedrun Ethereum](https://speedrunethereum.com) as they ship.

packages/nextjs/package.json

@@ -21,8 +21,10 @@
     "@uniswap/sdk": "^3.0.3",
     "daisyui": "^2.31.0",
     "ethers": "^5.0.0",
+    "gray-matter": "^4.0.3",
     "luxon": "^3.3.0",
     "next": "^13.4.1",
+    "next-mdx-remote": "^6.0.0",
     "nextjs-progressbar": "^0.0.16",
     "react": "^18.2.0",
     "react-blockies": "^1.4.1",

packages/nextjs/pages/blog/[slug].tsx

@@ -0,0 +1,95 @@
+import Head from "next/head";
+import type { GetStaticPaths, GetStaticProps, NextPage } from "next";
+import { MDXRemote, MDXRemoteSerializeResult } from "next-mdx-remote";
+import { serialize } from "next-mdx-remote/serialize";
+import { BlogMeta, getAllBlogSlugs, getBlogBySlug } from "~~/services/blog";
+
+const codeCls = "bg-white/5 text-primary-content px-1 py-0.5 rounded text-[0.85em] font-mono";
+
+// eslint-disable-next-line @typescript-eslint/no-explicit-any
+const components: Record<string, any> = {
+  h2: ({ children }: { children?: React.ReactNode }) => (
+    <h2 className="text-secondary font-bold text-lg sm:text-xl mt-10 mb-1">{children}</h2>
+  ),
+  a: ({ href, children }: { href?: string; children?: React.ReactNode }) => {
+    const isExternal = href?.startsWith("http");
+    return (
+      <a
+        href={href}
+        className="link link-primary"
+        {...(isExternal ? { target: "_blank", rel: "noopener noreferrer" } : {})}
+      >
+        {children}
+      </a>
+    );
+  },
+  code: ({ children }: { children?: React.ReactNode }) => <code className={codeCls}>{children}</code>,
+  strong: ({ children }: { children?: React.ReactNode }) => (
+    <strong className="text-white font-semibold">{children}</strong>
+  ),
+  ul: ({ children }: { children?: React.ReactNode }) => <ul className="mt-4 space-y-3 list-none">{children}</ul>,
+  li: ({ children }: { children?: React.ReactNode }) => (
+    <li className="flex items-start gap-3">
+      <span className="text-primary-content/40 font-mono shrink-0 mt-0.5">—</span>
+      <span>{children}</span>
+    </li>
+  ),
+};
+
+interface Props {
+  source: MDXRemoteSerializeResult;
+  meta: BlogMeta;
+}
+
+const BlogPost: NextPage<Props> = ({ source, meta }) => {
+  return (
+    <>
+      <Head>
+        <title>{meta.title} — Sand Garden</title>
+        <meta name="description" content={meta.description} />
+      </Head>
+
+      <article className="max-w-2xl px-4 py-12 mx-auto">
+        {/* Header */}
+        <header className="mb-10">
+          <div className="flex items-center gap-3 mb-4 font-mono text-sm text-white/35">
+            <span>{meta.date}</span>
+          </div>
+          <h1 className="text-3xl sm:text-4xl font-bold leading-tight text-white">{meta.title}</h1>
+        </header>
+
+        {/* Body */}
+        <div className="space-y-5 text-white/70 leading-relaxed text-sm sm:text-base">
+          <MDXRemote {...source} components={components} />
+        </div>
+
+        {/* Footer */}
+        <div className="mt-16 pt-6 border-t border-white/10 font-mono text-sm text-white/35">
+          <p>
+            Questions or want to build together?{" "}
+            <a href="mailto:sandgarden@buidlguidl.com" className="link link-primary">
+              sandgarden@buidlguidl.com
+            </a>
+          </p>
+        </div>
+      </article>
+    </>
+  );
+};
+
+export const getStaticPaths: GetStaticPaths = async () => {
+  const slugs = getAllBlogSlugs();
+  return {
+    paths: slugs.map(slug => ({ params: { slug } })),
+    fallback: false,
+  };
+};
+
+export const getStaticProps: GetStaticProps<Props> = async ({ params }) => {
+  const slug = params?.slug as string;
+  const { meta, content } = getBlogBySlug(slug);
+  const source = await serialize(content);
+  return { props: { source, meta } };
+};
+
+export default BlogPost;