How abd-ul-hannan/mybot works
Project type and purpose
• MyBot is a backend-focused, modular multi-agent AI system written in Python.
• It orchestrates specialized “sub-agents” (Inspiration, Automation, Content Creation, Customer Service, Data Analysis) to handle different domains.
• It is not a traditional front-end/full-stack web app but does expose a CLI/web entrypoint (adk web) and interacts with external services (OpenAI APIs, BigQuery, news/place APIs).
File structure
• Top-level:
• .env for configuration
• README.md for documentation
• agents/ — core orchestrator (agent.py, prompt.py) and sub_agents/ containing domain-specific implementations
• tools/ — utility modules (memory management, place lookup, web search)
• Each sub-agent folder follows a pattern:
• agent.py — agent orchestration logic
• prompt.py — prompt templates
• domain-specific scripts (e.g., automation_agent.py, Data_Analysis_Insights_agent.py)
• No front-end code; likely a CLI/HTTP wrapper around the core agents.
README for additional insights
• Confirms the five specialized agents and their responsibilities.
• Mentions external dependencies (BigQuery, news/place APIs, OpenAI).
• Notes use of a virtual environment and environment variables.