SDK for Mitii Core
Overview: A robust toolkit for integrating intelligent, context-aware AI agents into your development workflow.
- Core Tooling Interface: Provides the foundational functions the LLM needs to interact with the environment, including scan_workspace(), read_file(), web_search(), and write_file().
- Plan Mode (Plan-and-Execute): Decouples high-level strategic reasoning from low-level execution. This allows the agent to handle complex tasks reliably without getting confused or endlessly looping.
- Ask Mode (Interactive Approval): A safe, collaborative mode where the agent pauses to ask for confirmation rather than just running away with a task. The agent automatically generates an impact analysis and explicitly presents it to the user before changing any code.
- Agent Mode (Autonomous): The standard, basic prompt-and-response loop designed for quick tasks and immediate execution.
- State & Memory Management: Built-in conversational memory to handle multi-turn operations. This ensures the LLM successfully stores message history and remembers exactly what to execute after user approval.
SDK for Mitii Core
Overview: A robust toolkit for integrating intelligent, context-aware AI agents into your development workflow.