Enables AI agents to learn from their sessions by inspecting history, proposing rules or skills, and automatically persisting them to configuration files.
MCP server for agent-first Kanban board management, enabling agents to create, move, comment, and claim pebbles (cards) across statuses like requested, code_creation, testing, validation, and complete, with support for blocking, context feedback, and HIL (human-in-the-loop) transitions.