What is the 2-Action Rule mentioned in the documentation?

The 2-Action Rule mandates that after every two browser or search operations, findings must be written to disk to prevent visual or multimodal data from being lost.

When is it best to use this planning pattern?

It is recommended for any task requiring more than five tool calls, complex multi-step research, or projects where maintaining long-term context is difficult.

Where are the planning files stored?

Files are created in your current project directory (project root) so that they remain alongside your code and are easily accessible.

How does the 3-Strike Error Protocol help in development?

It prevents the AI from repeating the same failing action. It forces a diagnosis, followed by an alternative approach, then a broader rethink before escalating to the user.

What are the core files used by the Planning with Files skill?

The skill utilizes three primary files: task_plan.md for phases and decisions, findings.md for research storage, and progress.md for session logging.

Planning with Files

Name: Planning with Files
Author: claudiodearaujo

byclaudiodearaujo

생산성 및 워크플로

Organizes complex, multi-step AI tasks using persistent markdown-based planning files for enhanced context management and error recovery.

소개

This skill implements a sophisticated 'Manus-style' workflow that offloads Claude's short-term memory to project-level markdown files. By maintaining a living task_plan.md, findings.md, and progress.md, it ensures Claude stays on track during long-running tasks, captures critical insights from web searches, and follows a strict error-correction protocol. It is particularly effective for research projects, architectural changes, or any operation requiring more than five tool calls where context window drift becomes a risk to task success.

주요 기능

Context Refreshing Hooks: Pre-action hooks ensure the AI re-reads the plan before making major architectural decisions.
Manus-Style Execution: Implements a 'working memory on disk' pattern to prevent context loss during long sessions.
Persistent Task Tracking: Uses task_plan.md, findings.md, and progress.md to maintain project state on disk.
2-Action Research Rule: Automatically captures findings after every two search or browser operations to prevent data loss.
3-Strike Error Protocol: Enforces a structured diagnostic and escalation process for recurring technical failures.
0 GitHub stars

사용 사례

Automating multi-stage development workflows where state persistence across session boundaries is critical.
Conducting deep research tasks that require synthesizing information from various web sources into a coherent report.
Managing complex refactoring projects involving multiple files and cross-module dependencies.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add claudiodearaujo/sistema-de-narra-o-de-livro-front planning-with-files

For use in Claude.ai and ChatGPT

Download Skill

GitHub

소개

주요 기능

Context Refreshing Hooks: Pre-action hooks ensure the AI re-reads the plan before making major architectural decisions.
Manus-Style Execution: Implements a 'working memory on disk' pattern to prevent context loss during long sessions.
Persistent Task Tracking: Uses task_plan.md, findings.md, and progress.md to maintain project state on disk.
2-Action Research Rule: Automatically captures findings after every two search or browser operations to prevent data loss.
3-Strike Error Protocol: Enforces a structured diagnostic and escalation process for recurring technical failures.
0 GitHub stars

사용 사례

Automating multi-stage development workflows where state persistence across session boundaries is critical.
Conducting deep research tasks that require synthesizing information from various web sources into a coherent report.
Managing complex refactoring projects involving multiple files and cross-module dependencies.