How does the 3-strike error protocol work?

It forces a systematic approach to failures: first attempt is a diagnose and fix, second is an alternative method, and the third is a broader rethink before finally escalating to the user.

What is Manus-style planning?

It is a methodology that uses persistent Markdown files as 'working memory' to help AI agents manage complex, multi-step tasks without losing context as the conversation grows.

When should I use the planning-with-files skill?

This skill is recommended for any task requiring more than 5 tool calls, complex research, or building new features where maintaining a clear roadmap is essential.

Where are the planning files stored?

While the skill templates are stored in the plugin directory, your active planning files (task_plan.md, findings.md, and progress.md) are stored directly in your project's root directory.

Does it support session recovery?

Yes, it includes a session catch-up script that checks for unsynced changes and recovers context after a /clear command or a session reset.

Manus-Style Task Planning

Name: Manus-Style Task Planning
Author: amano--

byamano--

0•

Productivity & Workflow

Implements persistent, file-based orchestration for complex multi-step tasks to maintain long-term context and progress tracking.

This skill equips Claude with a structured, 'Manus-style' planning framework that uses local Markdown files as persistent working memory. By maintaining a task plan, findings log, and progress report directly in the project directory, it overcomes the volatility of the AI context window, making it ideal for research, building new features, or any project requiring more than five tool operations. It includes automatic session recovery and a strict error-handling protocol to ensure continuity and reliability during long-running development sessions.

Key Features

010 GitHub stars

02Automatic session recovery and context catch-up after clearing history

03Persistent state management using task_plan.md, findings.md, and progress.md

04Multimodal data persistence by saving web and browser findings to disk immediately

05Strict 3-Strike Error Protocol for systematic debugging and escalation

06Real-time phase tracking and Read Before Decide rule to prevent goal drift

Use Cases

01Large-scale refactoring tasks that span across many tool calls and sessions

02Complex feature implementation requiring multi-file edits and testing

03In-depth research projects involving multiple web searches and data synthesis

Key Features

010 GitHub stars

02Automatic session recovery and context catch-up after clearing history

03Persistent state management using task_plan.md, findings.md, and progress.md

04Multimodal data persistence by saving web and browser findings to disk immediately

05Strict 3-Strike Error Protocol for systematic debugging and escalation

06Real-time phase tracking and Read Before Decide rule to prevent goal drift

Use Cases

01Large-scale refactoring tasks that span across many tool calls and sessions

02Complex feature implementation requiring multi-file edits and testing

03In-depth research projects involving multiple web searches and data synthesis