This project demonstrates a fully functional autonomous AI agent built using Google’s Agent Development Kit (ADK) and powered by Gemini 2.5 Flash.
It simulates an intelligent multi-tool workflow — capable of researching, writing, and critiquing — showing how LLM-based agents can perform reasoning-driven, goal-oriented tasks.
The agent acts as a mini research assistant that:
- Collects relevant information (mock research tool),
- Synthesizes it into a coherent report (writer tool),
- Reviews the generated content (critic tool).
Agent Name: root_agent
Model Used: gemini-2.5-flash
| Stage | Tool | Purpose |
|---|---|---|
| 1️⃣ | research_tool(topic) |
Gathers relevant snippets about a user topic. |
| 2️⃣ | writer_tool(context) |
Generates a structured Markdown report using the snippets. |
| 3️⃣ | critic_tool(report, context) |
Evaluates coverage and coherence of the report. |