micro-fix :added database initialization and statement management tools

[IBOCATA]

Added database initialization and statement management tools.

Description

This PR introduces a comprehensive enhancement for the Hive AI Agent, enabling the creation and management of a dedicated String Database. This feature automates the initiation, training, and validation of business statements, providing a structured workflow for LLM-assisted human judgment.

Type of Change

• Bug fix (non-breaking change that fixes an issue)
• New feature (non-breaking change that adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)

Related Issues

Fixes #3393

Changes Made

• ✨ Key Features
  String DB Orchestration: Implements a managed layer for storing and indexing raw business statements imported into the system.
  Lifecycle Management: Automates the transition of statements through five distinct phases: Initiation, Training, Testing, Validation, and Final Judgment.
  Human-in-the-loop (HITL) Interface: Optimized outputs specifically designed for human auditors to review LLM "judgments" on business logic.
  Integrated Search Cache: Utilizes the previously implemented TTL cache to verify business statements against real-world data without redundant API costs.

Implementation

1. Database Initiation & Schema
   We use the Zod validation layer to ensure every imported business statement is "clean" before it hits the String DB.
2. The Training & Testing Pipeline
   The agent uses a "Reasoning Loop" to train on the business context and test the logic of the statements.
   Initiation: Raw CSV/JSON import of business statements.
   Training: LLM creates embeddings and associations within the String DB.
   Testing: Automated edge-case generation to stress-test statement logic.
   Validation: Comparing statement outcomes against cached web search results.

Testing

​🚦 Testing & Validation
​[x] Import Integrity: Verified that malformed business statements are rejected by the Zod gateway.
​[x] State Persistence: Confirmed that statements maintain their "Training" or "Testing" status across agent restarts.
​[x] Human Judgment Export: Verified the "Human-readable" summary format for final validation.

Checklist

• My code follows the project's style guidelines
• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes

📋 Assignee Checklist (SQLite/Vector Implementation)
Initialization: Ensure init_db() is called on server startup so the .db file is created locally.
Persistence Check: Verify that even if the MCP server restarts, the "Training" status of statements is preserved in SQLite.
Human Judgement Flow: The get_training_batch tool should be used to feed data to the LLM for initial "Human-like" evaluation