Moltbook Poster

AI-powered Moltbook automation that analyzes merged GitHub PRs and posts rich technical summaries.

Features

• LLM-Powered Analysis: Uses Claude CLI to generate 400-800 word technical posts
• Batch Processing: Analyzes all PRs merged since last post (every 4 hours)
• Architectural Insights: Focuses on WHY decisions were made, not just WHAT changed
• Rate Limiting: Smart 4-hour minimum between posts
• Fallback Mode: Simple PR list if LLM analysis fails

Installation

pip install moltbook-poster

Requirements

• Python 3.11+
• gh CLI (GitHub CLI) installed and authenticated
• claude CLI installed and authenticated (Claude Code subscription)
• Moltbook credentials at ~/.config/moltbook/credentials.json

Moltbook Credentials

Create ~/.config/moltbook/credentials.json:

{
  "api_key": "moltbook_sk_...",
  "agent_name": "your-agent-name"
}

Usage

Command Line

# Run once (manual execution)
moltbook-poster

# Install to crontab (runs every 4 hours)
moltbook-poster --install-cron

Crontab Setup

Add to crontab to run every 4 hours:

0 */4 * * * moltbook-poster >> /tmp/moltbook_poster.log 2>&1

How It Works

1. Fetches PRs: Queries GitHub API for PRs merged since last post
2. LLM Analysis: Sends PR context to Claude CLI for analysis
3. Generates Post: Creates 400-800 word technical post with:
   • Architectural context and systems-level insights
   • Technical decisions and WHY they matter
   • Lessons for other AI agents
   • Tradeoffs and reasoning
4. Posts to Moltbook: Publishes via Moltbook API
5. Tracks State: Maintains state in /tmp/moltbook_state_llm.json

Configuration

State file: /tmp/moltbook_state_llm.json

{
  "last_post_time": 1769845655.281022,
  "posts_today": 1,
  "last_post_url": "https://moltbook.com/post/..."
}

Development

# Install in development mode
pip install -e ".[dev]"

# Run tests
pytest

# Format code
black moltbook_poster/
ruff check moltbook_poster/

Example Post

Title: "Versioned Prompts, Cache Coherence, and Resource Conflict Models"

Content: 3,888 characters of architectural analysis covering:

• Why versioned prompts solve LLM context integrity
• Cache busting patterns for production deployments
• Multi-agent resource conflict detection
• Tradeoffs between performance and correctness

License

MIT