Can I modify the execution plan before it starts?

Yes. After the skill parses your task and presents a 3-7 step plan, you can choose to approve it, reject it, or provide feedback to modify the steps.

How does Step-by-Step ensure I remain in control of the code?

The skill is hardcoded to never proceed without explicit approval. It uses the AskUserQuestion tool to pause after generating a plan and after every single step, requiring your 'Yes' to continue.

What happens if a specific step fails during execution?

If a tool error occurs, the skill will report the failure details and prompt you via the AskUserQuestion tool to either retry the step, skip it, or abort the entire task.

Is there a limit to how many steps a task can have?

While it aims for 3-7 steps for clarity, it can handle larger tasks. However, if a task exceeds 20 steps, the skill will warn you and suggest breaking it into smaller, more manageable sub-tasks.

Interactive Step-by-Step Task Runner

Name: Interactive Step-by-Step Task Runner
Author: emirrtopaloglu

byemirrtopaloglu

0•

Productivity & Workflow

Executes complex tasks through atomic steps with mandatory user approval at every stage to ensure maximum precision.

The Step-by-Step skill transforms Claude into a highly supervised orchestrator, ideal for complex refactors or multi-stage implementations where granular control is paramount. It automatically decomposes high-level instructions into a structured plan of atomic actions, utilizing the AskUserQuestion tool to maintain a strict human-in-the-loop workflow. This ensures that every file change, command execution, or architectural decision is vetted by the developer before proceeding, preventing autonomous errors and providing clear progress tracking from start to finish.

Key Features

01Granular error handling with options to retry, skip, or abort on failure

02Seamless integration with specialized agents like @backend-architect or @frontend-architect

03Mandatory human-in-the-loop validation using the AskUserQuestion tool

040 GitHub stars

05Automatic task decomposition into 3-7 manageable, atomic steps

06Real-time progress tracking with status indicators for each completed stage

Use Cases

01Performing risky refactoring tasks where every change needs manual verification

02Orchestrating multi-step environment setups or database migrations with high oversight

03Implementing complex authentication flows that span multiple files and layers

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add emirrtopaloglu/claude-code-virtuoso step-by-step

For use in Claude.ai and ChatGPT

Download Skill