Add V4 Tool

This skill provides a comprehensive workflow for developers looking to expand the capabilities of the Agent V4 system. It outlines the technical requirements for defining, registering, and integrating new tool functions within the existing multi-agent architecture. The primary target audience includes backend engineers and AI developers working on the Pi Crypto Insight platform who need to introduce new functional modules for market data processing, social interaction, or internal system governance.

• Define tools as standard Python functions in core/agents/tools.py using proper type hints and descriptive docstrings to facilitate LLM function selection.

• Register functions via the ToolRegistry within core/agents/bootstrap.py, including metadata configuration for schema definitions and agent-specific access restrictions.

• Update AgentMetadata to ensure the Manager agent effectively routes tasks to the new capabilities, maintaining synchronization between tool availability and agent profiles.

• Utilize the internal self._use_tool interface for agent-to-agent communication and task execution, ensuring seamless logic flow within the LangGraph architecture.

• Ensure all tool parameters include explicit type hints to allow for accurate input schema validation by the system.

• Use descriptive docstrings as they serve as the primary source of truth for the LLM to determine tool suitability for specific user queries.

• Verify new implementations using the provided bootstrap diagnostic command to check tool registration status across the registry.

• Validate that custom tools adhere to the security and performance standards outlined in the broader platform documentation before integration into the production environment.

• The system enforces a strict separation between tool definitions and agent logic, requiring developers to align both the ToolMetadata and AgentMetadata registries to guarantee feature accessibility.