Meta-Prompting Framework

Recursive prompt improvement with real LLM integration

Transform AI outputs from good to great through recursive improvement

---

What Is This?

A real, working meta-prompting engine that recursively improves LLM outputs by:

1. Calling the LLM with an initial prompt
2. Extracting patterns and context from the response
3. Generating an improved prompt using that context
4. Repeating until quality threshold met

Not a simulation. Real Claude API calls with measurable improvements.

---

Key Features

• Recursive Improvement Loop - Automatic prompt refinement based on extracted context
• Complexity-Adaptive Strategies - Simple/Medium/Complex task routing
• Real Claude API Integration - Production-ready with token tracking
• Quality Threshold Control - Stop when target quality reached
• Claude Code Plugin - Install as skills, agents, and workflows
• Full Transparency - View all API calls in call_history

---

Proven Results

From our latest test with real Claude Sonnet 4.5:

Task: "Write function to find max number in list with error handling"

6 real API calls | 3,998 tokens | 89.7 seconds
- 2 complete iterations with context extraction
- Production-ready code with comprehensive error handling
- Full test suite included
- Two implementation variants (strict + lenient)

Typical improvement: 15-20% quality gain across iterations.

---

Quick Start

Option 1: Claude Code Plugin (Recommended)

git clone https://github.com/manutej/meta-prompting-framework.git
cd meta-prompting-framework
./install-plugin.sh
export ANTHROPIC_API_KEY=sk-ant-your-key-here

Option 2: Python Library

git clone https://github.com/manutej/meta-prompting-framework.git
cd meta-prompting-framework
pip install -r requirements.txt

from meta_prompting_engine.llm_clients.claude import ClaudeClient
from meta_prompting_engine.core import MetaPromptingEngine

engine = MetaPromptingEngine(ClaudeClient(api_key="your-key"))
result = engine.execute_with_meta_prompting(
    skill="python-programmer",
    task="Create a function to validate email addresses",
    max_iterations=3,
    quality_threshold=0.90
)
print(f"Quality: {result.quality_score:.2f}")

See docs/QUICKSTART.md for detailed setup instructions.

---

How It Works

Complexity	Strategy	Approach
< 0.3 (Simple)	Direct Execution	Single-pass with clear reasoning
0.3-0.7 (Medium)	Multi-Approach	Generate 2-3 approaches, evaluate, choose best
> 0.7 (Complex)	Autonomous Evolution	Hypothesize, test, refine iteratively

The engine extracts patterns, requirements, and success indicators from each iteration to build progressively better prompts.

---

Documentation

Document	Description
QUICKSTART.md	5-minute setup guide
INSTALL.md	Detailed installation options
PLUGIN_README.md	Claude Code plugin documentation
QUICK_REFERENCE.md	API reference card
CHANGELOG.md	Version history

Additional Resources

• docs/guides/ - Integration guides and workflow patterns
• docs/research/ - Research synthesis and references
• docs/internal/ - Implementation plans and summaries

---

Architecture

meta_prompting_engine/
  core.py           # MetaPromptingEngine - recursive loop
  complexity.py     # ComplexityAnalyzer - 0.0-1.0 scoring
  extraction.py     # ContextExtractor - 7-phase extraction
  llm_clients/
    base.py         # Abstract interface
    claude.py       # Claude Sonnet integration

---

Testing

# Validate without API key (uses mocks)
python3 validate_implementation.py

# Test with real Claude API
python3 test_real_api.py

# Show actual Claude responses
python3 show_claude_responses.py

---

License

MIT - see LICENSE file

---

Support

• Issues: GitHub Issues
• Docs: See docs/ directory

---

Built with real meta-prompting, not simulations.

Recursive improvement for better AI outputs.