memory-checkpoint

Externalize detailed memory to Memory Bank files before conversation summarization.

Updated Dec 19, 2025

Search Tags:#progress tracking #decision logging #context saving #memory preservation

Ready to run

npx skills add https://github.com/u9401066/copilot-capability-manager --skill memory-checkpoint

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: memory-checkpoint
Source: https://github.com/u9401066/copilot-capability-manager/tree/main/.claude/skills/memory-checkpoint
Command: npx skills add https://github.com/u9401066/copilot-capability-manager --skill memory-checkpoint

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill prevents critical project context from being lost during conversation summarization by proactively saving detailed memory to external files.

Core Features & Use Cases

Context Preservation: Externalizes detailed memory to Memory Bank before compression.
Use Case: When working on complex features or before taking breaks, use this Skill to capture all important decisions and progress.

Quick Start

Create a memory checkpoint to save the current conversation context and progress.

Frequently Asked Questions about memory-checkpoint

High-intent search queries and answers about installing and using this skill.

FAQPage Schema

How do I prevent losing context during long conversations or before taking a break?▼

Context preservation saves detailed memory to external files before conversation summarization. Create a checkpoint to batch-update memory files (activeContext, progress, decisionLog, architect) and retain all decisions and project state across long sessions or interruptions.

When should I use memory checkpoints in my workflow?▼

Use checkpoints after completing major features, during long-running sessions (over 10 turns), or manually trigger them with 'checkpoint', '存檔', or '記一下'. This prevents critical project context from being lost during conversation summarization.

What gets saved when I create a memory checkpoint?▼

A checkpoint externalizes detailed memory to Memory Bank by batch-updating four memory files: activeContext (current state), progress (completed work), decisionLog (choices made), and architect (design notes), while preserving file paths and decisions.

Can I trigger checkpoints automatically based on conversation length?▼

Yes, checkpoints are applied automatically during long-running sessions exceeding 10 turns or triggered manually via command. This ensures memory preservation happens consistently without manual intervention at regular intervals.

Do I need external memory bank setup before using checkpoints?▼

The Skill requires a Memory Bank as the storage destination for checkpoint files. Set up your memory file structure (activeContext, progress, decisionLog, architect) before creating checkpoints to batch-update all required files successfully.