jean-claude-dev

This skill provides a comprehensive development guide for maintaining and evolving the Jean Claude framework. It acts as a specialized assistant for contributors and maintainers, containing institutional knowledge, architectural patterns, and testing protocols required to modify the framework's source code, core orchestration logic, and system infrastructure. It is designed to bridge the gap between abstract project goals and concrete implementation details, ensuring consistency across complex AI-driven workflows. Use this skill when conducting deep-dive development tasks that involve the Jean Claude core SDK, event-sourced architecture, or workflow orchestration components.

• Detailed architectural guidance on the Two-Agent Workflow (Opus and Sonnet collaboration), SQLite-based event store patterns, and asynchronous mailbox communication between agents and coordinators.

• Implementation standards for Jean Claude development, including Git worktree isolation strategies and ntfy.sh escalation mechanisms for agent-to-human coordination.

• Comprehensive testing patterns, including mandatory mock patching rules (patching at the usage site versus definition site), proper fixture hierarchy usage, and AsyncMock conventions for async functions.

• Standardized documentation workflows, including retrospective logging, technical decision records (ADRs), and research integration documentation for new features.

• Guidance on running project-specific diagnostic tools, dashboard monitoring (SSE streaming), and memory leak identification within the agent orchestration pipeline.

• Reserved exclusively for contributors and maintainers working on the Jean Claude codebase or its internal architecture; do not invoke this for general end-user CLI interactions or simple workflow execution.

• Requires familiarity with the project structure, specifically src/jean_claude/ orchestration, core persistence, and CLI command modules.

• Emphasizes project-specific principles like testable business logic (avoiding over-mocking of Pydantic models or internal CLI frameworks) and safe parallel test execution.

• Integrates with the project's knowledge management system (plans, research, and decisions folders) to ensure that code changes align with documented architectural intent.

• Follows strict quality constraints such as mandatory ABOUTME file headers and the use of centralized fixture management in conftest.py files to minimize technical debt and redundant code patterns.