Where does this skill save the planning files?

While the skill templates stay in the skill folder, your specific task_plan.md, findings.md, and progress.md files are created and updated directly in your project root.

What is the 3-Strike Error Protocol?

It is a systematic debugging logic: Strike 1 is a targeted fix, Strike 2 is an alternative approach/tool, and Strike 3 is a complete rethink. After 3 failures, the AI escalates to the user.

Why use file-based planning instead of just chat context?

Chat context is volatile and limited (like RAM). Writing plans and findings to disk creates a persistent memory that won't be lost if the context window overflows or the session restarts.

When is it appropriate to use the planning-with-files skill?

Use it for any task requiring more than 5 tool calls, research-heavy projects, or building complex features. It's overkill for simple questions or single-file edits.

Manus-Style Task Planning

Name: Manus-Style Task Planning
Author: OthmanAdi

byOthmanAdi

•

11,606

생산성 및 워크플로

Implements a persistent, file-based workflow for Claude Code to manage complex, multi-step tasks with structured plan, findings, and progress logs.

소개

This skill provides Claude Code with a 'working memory on disk' by establishing a structured markdown-based planning system. Based on the high-performance workflow patterns popularized by AI agents like Manus, it directs the model to document goals, research discoveries, and session logs in local project files. This approach prevents context loss during long operations, enforces a rigorous 3-strike error protocol for debugging, and ensures project objectives remain prioritized by requiring the model to re-sync with the plan before making major architectural decisions.

주요 기능

Persistent file-based memory using task_plan.md, findings.md, and progress.md
Rigorous 3-Strike Error Protocol for systematic debugging and escalation
Context management decision matrix to optimize attention window usage
Pre-built automation scripts for session initialization and completion verification
11,606 GitHub stars
The 2-Action Rule to prevent multimodal and visual information loss

사용 사례

Conducting deep-dive research tasks where findings must persist across long sessions
Executing complex, multi-step software development projects requiring high tool-call volume
Managing large-scale refactors where tracking progress and error logs is critical

소개

주요 기능

Persistent file-based memory using task_plan.md, findings.md, and progress.md
Rigorous 3-Strike Error Protocol for systematic debugging and escalation
Context management decision matrix to optimize attention window usage
Pre-built automation scripts for session initialization and completion verification
11,606 GitHub stars
The 2-Action Rule to prevent multimodal and visual information loss

사용 사례

Conducting deep-dive research tasks where findings must persist across long sessions
Executing complex, multi-step software development projects requiring high tool-call volume
Managing large-scale refactors where tracking progress and error logs is critical