This is a structured approach to feature development that breaks work into five sequential phases: research, requirements, design, tasks, and implementation. You start with `/ralph-specum:start` and it walks you through creating markdown artifacts for each phase before autonomous execution begins. The quick mode skips the interviews and runs straight through. What's interesting is the epic triage system for large features, it decomposes them into smaller dependency-aware specs and tracks them in a shared state file. If you like planning before coding and want Claude to follow a consistent methodology instead of just starting to write code, this gives you that structure. The phase artifacts live in `./specs/<name>/` so you have a paper trail of decisions.

GitHub stars
340
last updated
May 19, 2026

Install

$ npx -y skills add tzachbon/smart-ralph --skill spec-workflow --agent claude-code

Installs into .claude/skills of the current project.

Files

Spec Workflow

Spec-driven development transforms feature requests into structured specs through sequential phases, then executes them task-by-task.

Decision Tree: Where to Start

SituationCommand
New feature, want guidance/ralph-specum:start <name> <goal>
New feature, skip interviews/ralph-specum:start <name> <goal> --quick
Large feature needing decomposition/ralph-specum:triage <goal>
Resume existing spec/ralph-specum:start (auto-detects)
Jump to specific phase/ralph-specum:<phase>

Single Spec Flow

start/new -> research -> requirements -> design -> tasks -> implement
                         ^
                         optional prototype overlay, then return

Each phase produces a markdown artifact under the resolved <basePath>/. Normal mode pauses for approval between phases. Quick mode runs all phases then auto-starts execution.

Sponsors

Details

Repository
github.com/tzachbon/smart-ralph
Categories
Data Science & ML, Automation & Workflows, Design & UI/UX
License
MIT
First seen
Jun 11, 2026

More from tzachbon/smart-ralph

All 28
1vila-lab avatarfigmirrorvila-lab/figmirrorThis FigMirror skill should be used when the user asks to "mirror this figure's style", "copy this figure's style", "make a chart that looks like this paper", "0skills.sh installs
2zhixiangluo avatarcreate-workflowzhixiangluo/10xproductivityDesign and build automation workflows using building blocks — clarify outcomes, decompose by dependencies, reuse prior art, verify each block before chaining, a0skills.sh installs
3huggingface avatarhugging-face-trackiohuggingface/skillsTrack and visualize ML training experiments with Trackio. Use when logging metrics during training (Python API), firing alerts for training diagnostics, or retr226skills.sh installs
4openai avatarjupyter-notebookopenai/skillsUse when the user asks to create, scaffold, or edit Jupyter notebooks (`.ipynb`) for experiments, explorations, or tutorials; prefer the bundled templates and r5.9Kskills.sh installs
5zc277584121 avatarjupyter-notebook-writingzc277584121/marketing-skillsWrite Milvus application-level Jupyter notebook examples using a Markdown-first workflow with jupyter-switch for format conversion.3.9Kskills.sh installs
6mindrally avatardeep-learning-pytorchmindrally/skillsExpert guidance for deep learning, transformers, diffusion models, and LLM development with PyTorch, Transformers, Diffusers, and Gradio.1.9Kskills.sh installs