How does the 3-strike error protocol work?

It prevents the AI from repeating mistakes by forcing a diagnostic fix on attempt 1, an alternative approach on attempt 2, and a broader rethink on attempt 3.

When is it best to use Structured File-Based Planning?

It is recommended for any task requiring more than 5 tool calls, complex research, or multi-step engineering work where keeping the goal in focus is critical.

What are the primary files used by this skill?

The skill creates and maintains task_plan.md (phases/decisions), findings.md (research storage), and progress.md (session logging) in your project directory.

Where are the planning files stored?

All planning files are created in your current project directory (CWD), ensuring they live alongside your code and are not buried in the skill's installation folder.

How does this solve context window limitations?

By treating the context window as RAM and the filesystem as Disk, it offloads important information to files so Claude can 're-read' the state if the conversation context becomes stale.

Structured File-Based Planning

Name: Structured File-Based Planning
Author: ninehills

byninehills

•

Productivity & Workflow

Implements a persistent, Manus-style file-based memory system for managing complex, multi-step engineering and research tasks.

About

This skill equips Claude with a robust project management framework that uses the filesystem as an extended 'working memory.' By creating and maintaining dedicated markdown files—task_plan.md, findings.md, and progress.md—the skill ensures that goals, research discoveries, and session logs persist even as the conversation grows. This approach prevents context window fatigue, minimizes repetitive errors through a structured 3-strike protocol, and provides a clear audit trail for complex software development or deep research projects spanning many tool calls.

Key Features

Integrated 3-strike error protocol to prevent repetitive logic failures and loops.
The 2-Action Rule for immediate preservation of research and multimodal data to text files.
Automatic pre-tool hooks that display the current plan to maintain context and focus.
5 GitHub stars
Persistent 'Memory on Disk' via structured task_plan.md, findings.md, and progress.md files.
Post-tool verification scripts to ensure phases are updated and completion criteria are met.

Use Cases

Long-running debugging sessions that involve multiple diagnostic attempts and error logging.
Deep research tasks where web data must be synthesized and stored outside volatile context.
Multi-step software engineering projects requiring complex state management.

About

Key Features

Integrated 3-strike error protocol to prevent repetitive logic failures and loops.
The 2-Action Rule for immediate preservation of research and multimodal data to text files.
Automatic pre-tool hooks that display the current plan to maintain context and focus.
5 GitHub stars
Persistent 'Memory on Disk' via structured task_plan.md, findings.md, and progress.md files.
Post-tool verification scripts to ensure phases are updated and completion criteria are met.

Use Cases

Long-running debugging sessions that involve multiple diagnostic attempts and error logging.
Deep research tasks where web data must be synthesized and stored outside volatile context.
Multi-step software engineering projects requiring complex state management.