Html Parser Added

Author: Saurabh Singh

CLAUDE.md

@@ -0,0 +1,77 @@
+# CLAUDE.md
+
+This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository.
+
+## Project Overview
+
+**promptplay** is a tool that uses the [Claude Agent SDK](https://github.com/anthropics/anthropic-sdk-python) to transform natural language prompts into fully functional, playable games. The agent receives a game description, uses Claude to write complete HTML/JavaScript game code, and outputs the result.
+
+## Setup & Common Commands
+
+### Initial Setup
+```bash
+# Create and activate virtual environment
+python -m venv venv
+source venv/bin/activate  # Windows: venv\Scripts\activate
+
+# Install dependencies
+pip install -r requirements.txt
+
+# Configure API key
+cp .env.example .env
+# Edit .env and add your ANTHROPIC_API_KEY
+```
+
+### Running the Agent
+```bash
+# Generate a game from the current prompt in agent.py
+python agent.py
+
+# Redirect output to an HTML file and open it
+python agent.py > my_game.html && open my_game.html
+```
+
+### Code Quality
+```bash
+# Check for import errors (run by CI)
+python -c "import agent; print('✓ All imports OK')"
+
+# Lint with ruff (only checks critical issues: syntax, undefined names, etc.)
+pip install ruff
+ruff check . --select=E9,F63,F7,F82 --show-source
+```
+
+## Architecture
+
+The project is minimal and focused:
+
+- **agent.py**: Single entry point using the Claude Agent SDK
+  - Uses `query()` async function with ClaudeAgentOptions
+  - Configurable model, allowed_tools, and effort level
+  - Streams responses and prints the result (typically HTML game code)
+  - Currently set to Haiku model for quick iteration
+
+- **Key Dependencies**:
+  - `claude-agent-sdk`: Provides the `query()` function and agent orchestration
+  - `python-dotenv`: Loads ANTHROPIC_API_KEY from .env file
+
+## Development Workflow
+
+1. **To test a prompt**: Edit the `prompt` string in `agent.py`, then run `python agent.py`
+2. **To switch models**: Change the `model` parameter (e.g., `"claude-opus-4-8"` for complex games)
+3. **To enable more tools**: Add tools to `allowed_tools` list (e.g., `["Bash", "Glob", "Read", "Write"]`)
+4. **Before committing**: Run the import check and verify the code runs without errors
+
+## Testing & CI
+
+- CI runs on Python 3.10, 3.11, 3.12
+- Tests verify: imports work, ruff linting passes
+- No dedicated test suite yet (see CONTRIBUTING.md for good first issues)
+
+## Key Files
+
+- `agent.py` – Main agent logic (async query to Claude)
+- `requirements.txt` – Dependencies
+- `.env.example` – Template for API key configuration
+- `CHANGELOG.md` – Version history
+- `CONTRIBUTING.md` – Guidelines for contributors

agent.py

@@ -1,4 +1,5 @@
 import asyncio
+import re
 from dotenv import load_dotenv
 from claude_agent_sdk import query, ClaudeAgentOptions
 
@@ -7,16 +8,25 @@
 
 async def main():
     async for message in query(
-        prompt="Build me a snake game in single index.html",
+        prompt="Write code for a snake game and give me one single index.html code",
         options=ClaudeAgentOptions(
             model="claude-haiku-4-5-20251001",
             allowed_tools=["Bash", "Glob"],
-            effort="high",
+            effort="low",
         ),
     ):
 
         if hasattr(message, "result"):
-            print(message.result)
+            html = htmlParser(message.result)
+            print(html)
+
+def htmlParser(result: str) -> str:
+    """Extract HTML from markdown code blocks."""
+    match = re.search(r'```(?:html)?\s*(.*?)\s*```', result, re.DOTALL)
+    if match:
+        return match.group(1).strip()
+    return result.strip()
+
 
 
 asyncio.run(main())