NEXUS: Universal Systems Programming Assignment Solver

NEXUS (Neural EXecution and Understanding System) is an autonomous AI-powered framework designed to solve complex systems programming assignments in C with high correctness. It integrates formal verification, symbolic execution, and LLM-guided code synthesis.

Prerequisites

• Python 3.8+
• GCC (for compiling generated C code)
• Make (for build management)

Systems Programming Roadmap

NEXUS now includes a comprehensive Systems Programming Masterclass Roadmap (12 Weeks) to help you master the fundamentals of systems programming while using the framework.

• View the Roadmap: See ROADMAP.md for the full curriculum.
• CLI Access: Run python3 main.py --roadmap to display the roadmap directly in your terminal.

Getting Started

1. Clone the Repository

git clone https://github.com/Devansh292004/NEXUS.git
cd NEXUS

2. Run the Framework

You can run the framework using the provided main.py entry point.

Using a string specification:

python3 main.py --spec "REQ: Implement a concurrent queue. REQ: Must be fast."

Using a specification file:

python3 main.py --file path/to/assignment.txt

Default run (sample specification):

python3 main.py

Generated Outputs

After a successful run, the framework produces:

• main.c: The synthesized C source code.
• Makefile: A professional Makefile with Debug/Release/Test variants.
• Documentation (accessible via the OutputGuarantees module):
  • System Architecture Diagrams (Mermaid.js)
  • API Documentation
  • Formal Proof of Correctness
  • Comprehensive README

Running Tests

To verify the framework's internal logic and pipeline integrity, run the comprehensive test suite:

export PYTHONPATH=$PYTHONPATH:.
python3 -m unittest discover tests

Architecture

The framework follows a 12-stage pipeline:

1. Multi-Stage Analysis: Requirement decomposition and constraint extraction.
2. Domain Planning: Specialized strategies for Memory, Concurrency, and Networking.
3. Formal Verification: Path exploration (KLEE) and SMT solving (Z3).
4. Code Synthesis: Multi-agent LLM-guided C code generation.
5. Intelligent Testing: Property-based, metamorphic, and mutation testing.
6. Compliance Checking: Static analysis for forbidden functions and code style.
7. Build System: Automatic Makefile generation.
8. Output Guarantees: Certification and documentation.