fix: Anthropic-style banner, restore README title, natural-language skill

- Redesign banner SVG: warm dark palette, minimal layout, clean monospace
- Restore <h1>Agentify</h1> title in README, add flat-square badge style
- Rewrite self/agentify.md as natural-language Claude Code skill with
  "Common User Requests" table for conversational interaction

Author: koriyoshi2041

README.md

@@ -1,13 +1,16 @@
 <div align="center">
 
-<img src="assets/banner.svg" alt="Agentify — Agent Interface Compiler" width="800">
+<h1>Agentify</h1>
+<p><strong>Agent Interface Compiler</strong> — One command. Every agent speaks your product.</p>
+
+<img src="assets/banner.svg" alt="Agentify — OpenAPI to 9 agent interface formats" width="800">
 
 <br>
 
-<a href="https://www.npmjs.com/package/agentify-cli"><img src="https://img.shields.io/npm/v/agentify-cli?color=cb3837&logo=npm" alt="npm"></a>
-<a href="https://github.com/koriyoshi2041/agentify/actions"><img src="https://img.shields.io/github/actions/workflow/status/koriyoshi2041/agentify/ci.yml?logo=github" alt="CI"></a>
-<a href="https://github.com/koriyoshi2041/agentify/blob/main/LICENSE"><img src="https://img.shields.io/github/license/koriyoshi2041/agentify" alt="MIT License"></a>
-<a href="https://www.typescriptlang.org/"><img src="https://img.shields.io/badge/TypeScript-strict-3178C6?logo=typescript&logoColor=white" alt="TypeScript"></a>
+<a href="https://www.npmjs.com/package/agentify-cli"><img src="https://img.shields.io/npm/v/agentify-cli?color=cb3837&logo=npm&style=flat-square" alt="npm"></a>
+<a href="https://github.com/koriyoshi2041/agentify/actions"><img src="https://img.shields.io/github/actions/workflow/status/koriyoshi2041/agentify/ci.yml?logo=github&style=flat-square" alt="CI"></a>
+<a href="https://github.com/koriyoshi2041/agentify/blob/main/LICENSE"><img src="https://img.shields.io/github/license/koriyoshi2041/agentify?style=flat-square" alt="MIT License"></a>
+<a href="https://www.typescriptlang.org/"><img src="https://img.shields.io/badge/TypeScript-strict-3178C6?logo=typescript&logoColor=white&style=flat-square" alt="TypeScript"></a>
 
 </div>
 

assets/banner.svg

@@ -1,66 +1,39 @@
 ---
 name: agentify
-description: Transform any OpenAPI spec into 9 agent interface formats (MCP Server, CLAUDE.md, AGENTS.md, .cursorrules, Skills, llms.txt, GEMINI.md, A2A Card, CLI)
+description: Transform any OpenAPI spec into agent-consumable formats — MCP Server, CLAUDE.md, AGENTS.md, .cursorrules, Skills, llms.txt, GEMINI.md, A2A Card, CLI
 ---
 
-You are helping the user transform an OpenAPI specification into agent interface formats using Agentify.
+Help the user make their API accessible to AI agents using Agentify.
 
-## Steps
+## What You Should Do
 
-1. Ask the user for the OpenAPI spec URL or file path if not provided
-2. Determine which formats to generate (default: all 8 standard formats, CLI is opt-in)
-3. Run the transform command:
+When the user asks you to make their API agent-accessible, generate agent interfaces, set up MCP, or anything related to converting an API for agent consumption:
 
-```bash
-npx agentify-cli transform <input> [options]
-```
+1. **Find the OpenAPI spec** — ask the user for the path or URL. If they point to a project directory, look for `swagger.json`, `openapi.json`, `openapi.yaml`, or similar files.
 
-### Options
+2. **Run the transform**:
+   ```bash
+   npx agentify-cli transform <spec-path-or-url>
+   ```
 
-- `-o, --output <dir>` — Output directory (default: `<product-name>-mcp-server`)
-- `-n, --name <name>` — Override project name
-- `-f, --format <formats...>` — Select formats: mcp, claude.md, agents.md, cursorrules, skills, llms.txt, gemini.md, a2a, cli
+3. **Guide them through the output** — the tool generates a complete project directory with all agent interface formats. Walk them through `cd <output>`, `npm install`, configuring `.env`, and `npm start`.
 
-### Available Formats
+## Common User Requests
 
-| Format | Flag | What it generates |
-|--------|------|-------------------|
-| MCP Server | `mcp` | Full TypeScript server with tools, handlers, Dockerfile |
-| CLAUDE.md | `claude.md` | Project context for Claude Code |
-| AGENTS.md | `agents.md` | Universal agent instructions (Linux Foundation standard) |
-| .cursorrules | `cursorrules` | Cursor IDE agent rules |
-| Skills | `skills` | Structured capability file (skills.json) |
-| llms.txt | `llms.txt` | LLM-readable condensed documentation |
-| GEMINI.md | `gemini.md` | Gemini CLI project context |
-| A2A Card | `a2a` | Google Agent-to-Agent discovery card |
-| CLI | `cli` | Standalone command-line tool (opt-in, not in defaults) |
+| User says | What to do |
+|-----------|------------|
+| "Make my API work with Claude" | `npx agentify-cli transform <spec> -f mcp claude.md` |
+| "Generate agent interfaces for my product" | `npx agentify-cli transform <spec>` (all formats) |
+| "I need an MCP server for this API" | `npx agentify-cli transform <spec> -f mcp` |
+| "Create a CLI from my API spec" | `npx agentify-cli transform <spec> -f cli` |
+| "Set up agent access for Cursor/Copilot" | `npx agentify-cli transform <spec> -f agents.md cursorrules` |
 
-## Examples
+## Options
 
-```bash
-# All default formats
-npx agentify-cli transform https://petstore.swagger.io/v2/swagger.json
-
-# Only MCP server
-npx agentify-cli transform ./api-spec.yaml -f mcp -o my-mcp-server
-
-# Generate a CLI tool
-npx agentify-cli transform ./spec.json -f cli
-
-# Specific formats
-npx agentify-cli transform ./spec.json -f mcp skills a2a
-```
+- `-o, --output <dir>` — where to write output (default: `<name>-mcp-server`)
+- `-n, --name <name>` — override project name
+- `-f, --format <formats...>` — pick specific formats: mcp, claude.md, agents.md, cursorrules, skills, llms.txt, gemini.md, a2a, cli
 
 ## Supported Inputs
 
-- Swagger 2.0 (JSON/YAML)
-- OpenAPI 3.0.x / 3.1.x (JSON/YAML)
-- URL or local file path
-
-## After Generation
-
-Guide the user through next steps:
-1. `cd <output-dir>`
-2. `npm install`
-3. `cp .env.example .env` and configure API keys
-4. `npm start`
+Swagger 2.0 or OpenAPI 3.x, JSON or YAML, URL or local file path.

self/agentify.md

@@ -24,7 +24,7 @@ describe("Self-Description Files", () => {
       );
       const match = content.match(/description:\s*(.+)/);
       expect(match).not.toBeNull();
-      expect(match![1]).toContain("agent interface formats");
+      expect(match![1]).toContain("agent-consumable formats");
     });
 
     it("contains the transform command usage", () => {
@@ -48,13 +48,13 @@ describe("Self-Description Files", () => {
       }
     });
 
-    it("includes examples section", () => {
+    it("includes common user requests for natural interaction", () => {
       const content = fs.readFileSync(
         path.join(SELF_DIR, "agentify.md"),
         "utf-8",
       );
-      expect(content).toContain("## Examples");
-      expect(content).toContain("petstore.swagger.io");
+      expect(content).toContain("## Common User Requests");
+      expect(content).toContain("## What You Should Do");
     });
   });