Skip to content

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

🔍 Beyan — AI-Powered Agentic Project Analysis Framework

Autonomous Intelligence for Deep Technical Audits & Refactoring

version agentic languages license

English · Türkçe


🚀 The Agentic Evolution (v2.0)

Beyan has evolved from a static prompt library into a fully autonomous agentic framework. It no longer just provides "what" to ask; it manages the "how" of analysis and execution.

🧠 The Beyan v2.0 Engine

  • Autonomous Discovery: Automatically fingerprints 20+ technologies in your project.
  • Smart Compilation: Dynamically builds a context-dense master prompt by selecting only relevant expert modules.
  • Agentic Loop (Mode 3): A semi-autonomous cycle that follows a Analyze → Plan → Code → Test → Commit workflow with human-in-the-loop safety gates.
  • Token Budgeting: Intelligent module pruning to fit massive codebase context into LLM limits.

🚦 Quick Start

Get your first deep analysis report in seconds:

# 1. Setup
git clone https://github.com/XINMurat/beyan.git
cd beyan/v2
pip install -r requirements.txt

# 2. Run Analysis (Mode 1)
python cli/analyzer.py --target /path/to/your/project --mode 1 --lang en --api openai

# 3. Interactive Fix (Mode 3)
python cli/analyzer.py --target /path/to/your/project --mode 3 --lang en --api anthropic

⚙️ How It Works (The v2.0 Workflow)

Unlike the manual process of v1.0, Beyan v2.0 automates the entire analytical pipeline:

graph TD
    A[Target Project] --> B{Discovery Engine}
    B -- Identifies Tags --> C[Go, Node, K8s, Security, etc.]
    C --> D{Compiler}
    D -- Loads Modules --> E[Custom Compiled Prompt]
    E --> F{Agentic Loop}
    F -- Mode 1 --> G[Final Analysis Report]
    F -- Mode 2 --> H[Implementation Plan]
    F -- Mode 3 --> I[Interactive Code Refactoring]
Loading

🧩 The Core Knowledge Base (Legacy v1.0 Prompts)

Beyan v2.0 is powered by the battle-tested v1.0 prompt family. You can still use these prompts manually for specialized needs:

Project-Type Prompts

Focus & Tooling


🏛️ Design Principles

  • NOT DETECTED Contract: If information isn't found, the agent flags it. Never halluncinates.
  • Two-Layer Analysis: Descriptive layer (facts) first, then Evaluative layer (judgment).
  • Safety Gating: Mode 3 creates a safety branch and asks for confirmation before every single file change.
  • Reconstructibility: "A new developer should be able to reconstruct the entire system solely from the analysis output."

📂 Project Structure

  • v2/: The core Agentic Framework (CLI, Orchestrator, Discovery).
  • tr/ & en/: The multi-language expert prompt modules.
  • .github/: CI/CD workflows for automated testing.

📝 Articles

⚖️ License

MIT — free to use, modify, and distribute.


Acknowledgements

Beyan was developed through an iterative self-referential process: designed, then audited using its own Meta Audit prompt, improved based on findings, and health-scored before release.

About

AI-powered agentic framework for deep code analysis — auto-discovery, modular compilation & semi-autonomous refactoring (TR/EN)

Topics

Resources

Contributing

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages