A goal-driven Agentic AI system built using a planner–executor–controller architecture. The agent decomposes high-level goals into steps, executes them safely, and persists execution history using JSON-based memory.
Planner → Executor → Controller → Memory
- Planner: Converts goals into ordered steps
- Executor: Executes steps using safe, rule-based tools
- Controller: Manages execution flow
- Memory: Stores execution history across runs
- Modular agent design
- Safe execution with command whitelisting
- Persistent JSON-based memory
- LLM-agnostic planner design
python test_controller.py