Show HN: DonnyClaude – a verified workflow engine for Claude Code (github.com)

DonnyClaude, a new workflow engine for Claude Code, has been unveiled, addressing the limitations of AI-generated code in managing long-term projects. While Claude Code excels in code generation, it struggles with maintaining context, leading to incomplete work artifacts and unchecked project milestones. DonnyClaude introduces a robust planning layer and a deterministic verification loop that ensures project progress is durable, structuring tasks into clear phases and requiring verification through specialized subagents for coding, reviewing, and security checks. This means that work will not be marked as complete until all checks have been satisfied. For the AI/ML community, this innovation is significant as it tackles common pitfalls in AI-assisted coding, promoting accountability and reliability in project management, especially for larger endeavors. By establishing a distinct workflow architecture, DonnyClaude differentiates itself from typical Claude Code setups, which often rely on brute-force prompt modifications without standardized completion criteria. The engine's installation is streamlined via a single command, incorporating extensive tools such as planners, verifiers, and coding standards, thereby enhancing both collaboration and development efficiency while minimizing the risks associated with AI-generated outputs.