Unlocking Claude’s Instant Evolution with the Open‑Source “planning‑with‑files” Skill

The open‑source “planning‑with‑files” project reproduces Manus’s workflow, showing that AI’s long‑term stability hinges on external memory via Context Engineering, and provides a Claude Skill that structures tasks with markdown files to prevent forgetting and goal drift.

AI Insight Log
AI Insight Log
AI Insight Log
Unlocking Claude’s Instant Evolution with the Open‑Source “planning‑with‑files” Skill

Heavy users observe that after 20–50 dialogue turns an AI model begins to forget its original goal, repeat past errors, and overflow the context window; this is caused by the limited working memory of the model and results in goal drift.

“Markdown is my ‘working memory’ on disk. Because my active context is limited, markdown files serve as both a note board and a set of checkpoints for progress and final deliverables.” – OthmanAdi

The method treats markdown files as the AI’s hard‑disk memory. The Skill enforces a “three‑file mode”: task_plan.md (Task Plan) – records goals, current stage, progress, and encountered errors. notes.md (Notes/Draft) – stores research results, intermediate thoughts, and long text, keeping them out of the limited conversation context. [deliverable].md (Deliverable) – contains the final output.

A typical AI agent follows a linear chain that quickly loses direction:

Receive → Execute → Execute → … → Lost → Error → Crash

With the Skill installed, Claude follows a disciplined four‑step cycle:

Read : Load task_plan.md to bring the goal back into focus.

Act : Perform searches, coding, or research.

Update : Write results to notes.md and tick off items in task_plan.md.

Review : Record any errors for future reference.

This “Read‑Decision‑Act‑Update” loop lets the AI remain aware of its identity, location, obstacles, and next steps even after 50+ tool calls.

The approach mirrors the GTD (Getting Things Done) methodology: external storage (markdown files) replaces reliance on the model’s internal memory, making context management more reliable.

Installation

After ensuring Claude Code is installed, clone the repository into the Claude Skills directory. The Skill activates automatically when the user mentions keywords such as “planning”, “organize”, or “track progress”.

# Enter Skills directory
cd ~/.claude/skills
# Clone the repository
git clone https://github.com/OthmanAdi/planning-with-files.git

Once installed, the Skill creates task_plan.md and begins its orchestrated workflow.

Project URL: https://github.com/OthmanAdi/planning-with-files

Original Source

Signed-in readers can open the original source through BestHub's protected redirect.

Sign in to view source
Republication Notice

This article has been distilled and summarized from source material, then republished for learning and reference. If you believe it infringes your rights, please contactadmin@besthub.devand we will review it promptly.

AI AgentGitHubClaudemarkdownContext EngineeringSkill
AI Insight Log
Written by

AI Insight Log

Focused on sharing: AI programming | Agents | Tools

0 followers
Reader feedback

How this landed with the community

Sign in to like

Rate this article

Was this worth your time?

Sign in to rate
Discussion

0 Comments

Thoughtful readers leave field notes, pushback, and hard-won operational detail here.