Skip to content

README.md

Latest commit

 

History

History
49 lines (38 loc) · 1.94 KB

README.md

File metadata and controls

49 lines (38 loc) · 1.94 KB

POS Change Request Evaluation System

This project implements a Multi-Agent System to evaluate Change Requests (CRs) for a Point of Sale (POS) application.

🧠 Project Memory / Context

Goal: Automate the initial review of features/requests using AI agents representing different stakeholders (Product Owner, Architect, Security, etc.).

System Architecture

  • Orchestration Layers:
    1. Generation: generate_cr.py & generate_functional_spec.py (Create content).
    2. Evaluation: evaluate.py (Review content).
    3. Summarization: generate_director_brief.py (Strategic view).
  • Agents: Defined in agents/definitions.py (Evaluators) and agents/generation_agents.py (Creators).
  • Knowledge Base (RAG): knowledge_base/ containing policy docs.
  • Web handlers: web_handlers.py for future API integration.

Current Status (Version 2.0 - Jan 2026)

  • Full Lifecycle Support: From "Idea" -> "CR" -> "Spec" -> "Evaluation".
  • Context Aware: Agents are configured for C#/.NET Core & XML architecture.
  • Director-Ready: Dedicated scripts to generate high-level strategic briefs (_DIRECTOR_BRIEF.md).
  • Input Flexibility: CLI accepts raw text or .txt files.

🚀 How to Run

  1. Create a Request:

    python create_cr.py "My New Feature"

    This creates change_requests/my_new_feature/my_new_feature.md.

  2. Run Evaluation:

    python evaluate.py change_requests/my_new_feature/my_new_feature.md

  3. View Report: Open change_requests/my_new_feature/my_new_feature_REPORT.md to see the Director's Summary and Agent details.

Repository Structure

  • agents/: Agent personas.
  • change_requests/: Database of CRs (one folder per CR).
  • knowledge_base/: RAG context files.
  • templates/: Templates.

Future Plans

  • Integrate real LLM API.
  • Add PDF parsing support for legacy CRs.
  • Create a Web UI for submission.