centricconsulting/agent_c_framework

AI Summary: Replace the existing chunked HTTP streaming in the Agent C framework with a WebSocket connection for bidirectional event streaming. This involves modifying the client to establish a WebSocket connection upon initialization and updating the agent bridge to publish events directly to the WebSocket. The goal is to enable real-time, bi-directional communication and resolve issues like the Sophos batching problem.

AI Summary: The task requires developing a new tool for the Agent C framework to integrate with the Plane API. This tool will allow Agent C agents to interact with Plane as an issue tracking system, enabling features like creating, updating, and querying issues within Plane directly from the agent's workflow. The tool should be designed to minimize token consumption and leverage Agent C's existing tool architecture and event-driven communication.

AI Summary: Integrate the ProseMirror rich text editor into the Agent C framework's user interface, replacing existing chat and prompt input fields to enable features like formatting and structured editing.

AI Summary: Implement a user interface feature within the Agent C framework to handle user dialog events. This involves creating a mechanism for tools to display web forms (specifically, yes/no questions initially) to the user, receive their responses, and then relay that information back to the tool for further processing. The solution should involve creating a `UserDialogEvent` for sending requests and a `UserDialogEventResponse` for returning user input, managing asynchronous communication between the UI and the tool, and integrating this functionality into the existing event-driven architecture.

AI Summary: Define a user profile data schema for an AI agent framework, including skill levels for various topics. This involves determining appropriate skill levels (e.g., beginner, intermediate, expert) for different areas relevant to the agent's capabilities and designing a system to represent and utilize this information within the Agent C framework. The goal is to tailor agent interactions and responses based on the user's proficiency.

AI Summary: The task requires consolidating existing web search tools within the Agent C framework into a single, master tool. This master tool should allow users to specify which search engine(s) to utilize, mirroring the functionality of the existing workspace tool's selection of storage backends. This involves modifying the existing codebase to create a unified interface and potentially refactoring existing search tool code for better integration.

AI Summary: Implement a simplified work log system for the Agent C framework. This involves creating new models (`WorklogEntry` and potentially updating `InteractionContext`) to track tool calls with minimal information (tool class, call ID, description). The `Toolset` class needs a logging method (`_log_work`) to facilitate this. This work log will be used to progressively summarize past interactions for inclusion in system prompts, improving efficiency and reducing token consumption.

AI Summary: Implement a container class for managing chat messages in the Agent C framework. This class should handle translation to/from a common message format, inherit from an Observable base class, support retrieval of messages by interaction ID and time ranges, and include interaction ID in the message models. This is a foundational step for enabling UI features like message editing and deletion.