feat: migrate blog storage to Supabase

Replace file-based KV storage with Supabase (PostgreSQL) for blog
post persistence. Each CRUD operation issues a direct SQL query via
@supabase/supabase-js. Adds CI secrets for the new database env vars.

Required env vars: SUPABASE_URL, SUPABASE_SERVICE_ROLE_KEY

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: KebanFiru

.github/workflows/pipeline.yaml

@@ -9,6 +9,20 @@ on:
       - main
 
 jobs:
+  check-lint:
+    name: Check lint
+    runs-on: ubuntu-latest
+    steps:
+      - uses: actions/checkout@v6
+
+      - uses: actions/setup-node@v6
+        with:
+          node-version: 22
+
+      - run: npm install > /dev/null 2>&1
+
+      - run: npm run lint > /dev/null 2>&1
+
   unit-tests:
     name: Jest unit tests
     runs-on: ubuntu-latest
@@ -26,7 +40,7 @@ jobs:
   deploy:
     name: Deploy to Vercel
     runs-on: ubuntu-latest
-    needs: [unit-tests]
+    needs: [unit-tests, check-lint]
     if: github.ref == 'refs/heads/main' && github.event_name == 'push'
 
     steps:
@@ -44,5 +58,11 @@ jobs:
           VERCEL_PROJECT_ID: ${{ secrets.VERCEL_PROJECT_ID }}
 
       - run: vercel build --prod --token=${{ secrets.VERCEL_TOKEN }} > /dev/null 2>&1
+        env:
+          SUPABASE_URL: ${{ secrets.SUPABASE_URL }}
+          SUPABASE_SERVICE_ROLE_KEY: ${{ secrets.SUPABASE_SERVICE_ROLE_KEY }}
 
       - run: vercel deploy --prebuilt --prod --token=${{ secrets.VERCEL_TOKEN }} > /dev/null 2>&1
+        env:
+          SUPABASE_URL: ${{ secrets.SUPABASE_URL }}
+          SUPABASE_SERVICE_ROLE_KEY: ${{ secrets.SUPABASE_SERVICE_ROLE_KEY }}

.gitignore

@@ -36,8 +36,6 @@ pnpm-lock.yaml
 # vercel
 .vercel
 
-/data/posts.json
-
 # typescript
 *.tsbuildinfo
 next-env.d.ts

README.md

@@ -14,6 +14,8 @@ It is designed for education, experimentation, and fast visual iteration.
 - Turing Machines: deterministic and nondeterministic, single-tape and multi-tape transitions
 - Mealy Machines
 - Moore Machines
+- GNFA (Generalized NFA) — for FA → regex extraction
+- CFG (Context-Free Grammar) — skeleton extraction from saved automata
 
 ## What You Can Do
 
@@ -36,10 +38,10 @@ It is designed for education, experimentation, and fast visual iteration.
 
 ### Conversion and Generation Tools
 
-- Regex conversion tool for finite automata workflows
-- NFA to DFA conversion for selected FA models
-- FA to GNFA generation
-- CFG generation and grammar-oriented artifacts from selected models
+- Regex → NFA via Thompson's construction
+- NFA → minimal DFA via subset construction and minimization
+- FA → GNFA for regular expression extraction via state elimination
+- CFG skeleton generation from saved finite automata
 
 ### Simulation and Visual Feedback
 
@@ -54,6 +56,53 @@ It is designed for education, experimentation, and fast visual iteration.
 - Regex summary views for finite automata cards
 - Machine summaries (model-specific tuple/style summaries)
 
+## AI Accessibility — MCP Server
+
+The simulator exposes an HTTP [Model Context Protocol](https://modelcontextprotocol.io/) endpoint at `POST /api/mcp`. AI agents (Claude Code, Cursor, GPT, etc.) can use it to simulate and build automata without touching the canvas.
+
+The endpoint implements JSON-RPC 2.0 and exposes 17 tools:
+
+**Simulate**
+
+| Tool | Description |
+|------|-------------|
+| `simulate_fa` | Test a string against a DFA or NFA |
+| `simulate_pda` | Test a string against a deterministic PDA |
+| `simulate_ndpa` | Test a string against a nondeterministic PDA |
+| `simulate_tm` | Test a string against a single/multi-tape TM |
+| `simulate_mealy` | Run a Mealy machine, return output sequence |
+| `simulate_moore` | Run a Moore machine, return output sequence |
+
+**Convert**
+
+| Tool | Description |
+|------|-------------|
+| `regex_to_nfa` | Convert regex to NFA via Thompson's construction |
+| `nfa_to_dfa` | Determinize NFA → minimal DFA |
+| `fa_to_regex` | Extract regex from FA via GNFA elimination |
+| `get_transition_table` | Render any automaton as a transition table |
+| `get_automata_info` | Full reference: types, tuple definitions, syntax |
+
+**Canvas links** — returns a URL that opens the automaton directly on the simulator canvas
+
+| Tool | Machine |
+|------|---------|
+| `build_fa_link` | DFA / NFA |
+| `build_pda_link` | Deterministic PDA |
+| `build_ndpa_link` | Nondeterministic PDA |
+| `build_tm_link` | Single/multi-tape Turing Machine |
+| `build_mealy_link` | Mealy machine |
+| `build_moore_link` | Moore machine |
+
+**Quick test:**
+```bash
+curl -X POST https://www.theoryofcomputationsimulator.com/api/mcp \
+  -H "Content-Type: application/json" \
+  -d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'
+```
+
+An OpenAPI 3.1 spec is available at [`/openapi.json`](https://www.theoryofcomputationsimulator.com/openapi.json). Full reference for LLM agents is at [`/llms.txt`](https://www.theoryofcomputationsimulator.com/llms.txt).
+
 ## Transition Label Syntax Reference
 
 ### Finite Automata
@@ -118,21 +167,45 @@ It is designed for education, experimentation, and fast visual iteration.
 2. Reload later from the workspace loader.
 3. Continue editing with previous state and saved cards restored.
 
-## Technical Highlights
-
-- Strongly typed model layer for state and alphabet handling
-- Regex simplification utilities integrated with FA summary flows
-- Language-equivalence-aware conversion safeguards in finite automata paths
-- Shared parsing/codec utilities for pushdown and machine snapshots
-- Refactored common FA view helpers to reduce duplicated logic
+### 5) Use with an AI Agent
+
+Ask an AI agent to design an automaton. It calls `build_fa_link` and returns a URL that opens it on the canvas:
+
+```bash
+curl -X POST https://www.theoryofcomputationsimulator.com/api/mcp \
+  -H "Content-Type: application/json" \
+  -d '{
+    "jsonrpc": "2.0", "id": 1,
+    "method": "tools/call",
+    "params": {
+      "name": "build_fa_link",
+      "arguments": {
+        "states": ["q0","q1","q2"],
+        "alphabet": ["a","b"],
+        "transitions": [
+          {"from":"q0","to":"q1","symbol":"a"},
+          {"from":"q1","to":"q2","symbol":"b"},
+          {"from":"q2","to":"q2","symbol":"b"}
+        ],
+        "start_state": "q0",
+        "accept_states": ["q2"]
+      }
+    }
+  }'
+```
+
+The response includes a `url` field. Open it and the automaton appears on the canvas with states positioned automatically.
 
 ## Technology Stack
 
-- Next.js 16
+- Next.js 16 (App Router)
 - React 19
-- TypeScript 5
+- TypeScript 6 (strict mode)
 - Tailwind CSS 4
 - Lucide React
+- Biome (linting & formatting)
+- Vercel KV (blog post storage)
+- Jest 30 (unit tests — 231 tests across 15 suites)
 
 ## Notes
 

app/admin/login/page.tsx

@@ -1,6 +1,6 @@
 "use client";
-import React from "react";
 import { useRouter, useSearchParams } from "next/navigation";
+import React from "react";
 
 function LoginForm() {
   const router = useRouter();
@@ -24,24 +24,31 @@ function LoginForm() {
 
     if (res.ok) {
       const raw = params.get("from") ?? "";
-      const from = raw.startsWith("/") && !raw.startsWith("//") ? raw : "/admin";
+      const from =
+        raw.startsWith("/") && !raw.startsWith("//") ? raw : "/admin";
       router.push(from);
     } else {
       setError("Invalid password");
     }
   }
 
   return (
-    <form onSubmit={handleSubmit} style={{ display: "flex", flexDirection: "column", gap: 16 }}>
+    <form
+      onSubmit={handleSubmit}
+      style={{ display: "flex", flexDirection: "column", gap: 16 }}
+    >
       <div style={{ display: "flex", flexDirection: "column", gap: 6 }}>
-        <label style={{ fontSize: 12, fontWeight: 600, color: "var(--text-muted)" }}>
+        <label
+          htmlFor="admin-password"
+          style={{ fontSize: 12, fontWeight: 600, color: "var(--text-muted)" }}
+        >
           Password
         </label>
         <input
+          id="admin-password"
           type="password"
           value={password}
           onChange={(e) => setPassword(e.target.value)}
-          autoFocus
           required
           style={{
             padding: "10px 14px",
@@ -54,7 +61,9 @@ function LoginForm() {
           }}
         />
         {error && (
-          <p style={{ fontSize: 12, color: "var(--danger)", margin: 0 }}>{error}</p>
+          <p style={{ fontSize: 12, color: "var(--danger)", margin: 0 }}>
+            {error}
+          </p>
         )}
       </div>
 
@@ -95,8 +104,17 @@ export default function AdminLogin() {
         className="ui-panel"
         style={{ width: "100%", maxWidth: 360, padding: "40px 36px" }}
       >
-        <p className="ui-kicker" style={{ marginBottom: 8 }}>Admin</p>
-        <h1 style={{ fontSize: 22, fontWeight: 700, letterSpacing: "-0.02em", marginBottom: 28 }}>
+        <p className="ui-kicker" style={{ marginBottom: 8 }}>
+          Admin
+        </p>
+        <h1
+          style={{
+            fontSize: 22,
+            fontWeight: 700,
+            letterSpacing: "-0.02em",
+            marginBottom: 28,
+          }}
+        >
           Sign in
         </h1>