What does the Context Preservation skill do?

It prevents Claude from forgetting important details during long development sessions by marking critical information and creating structured task checkpoints.

When does this skill activate?

It activates automatically during long coding sessions or immediately before the AI context window undergoes compaction.

Can I manually save my progress?

Yes, you can use the /checkpoint command to trigger an immediate save of your current task state, completed steps, and next actions.

How do I mark information as important?

You can wrap vital text in tags to ensure it is prioritized and survives context compaction.

Where is my session memory stored?

The skill saves data locally to the .claude/memory/ directory in files such as attention.md and checkpoint.md.

Context Preservation System

Name: Context Preservation System
Author: TheGlitchKing

byTheGlitchKing

生産性とワークフロー

Maintains critical project context and session history through intelligent attention marking and automated checkpoints.

概要

The Context Preservation System for Claude Code ensures that vital information like architectural decisions, blocking issues, and specific configurations are never lost during long sessions or context compaction. By utilizing a specialized attention marker system and structured checkpointing, it creates a persistent memory layer in your local directory, allowing the AI to resume tasks seamlessly even after memory resets or context compression events. This skill is essential for complex, multi-stage development projects where maintaining a 'long-term memory' of decisions and environment states is critical for success.

主な機能

Attention Marker System for tagging critical architectural decisions
Persistent memory management via dedicated local markdown files
Quick recovery protocols for resuming work after context resets
Structured task checkpointing for current state and next steps
Automated pre-compaction checklists to prevent information loss
0 GitHub stars

ユースケース

Ensuring user preferences and project goals persist across session resets
Managing multi-day coding sessions where context window limits are reached
Tracking complex debugging paths and discovered environment configurations

概要

主な機能

Attention Marker System for tagging critical architectural decisions
Persistent memory management via dedicated local markdown files
Quick recovery protocols for resuming work after context resets
Structured task checkpointing for current state and next steps
Automated pre-compaction checklists to prevent information loss
0 GitHub stars

ユースケース

Ensuring user preferences and project goals persist across session resets
Managing multi-day coding sessions where context window limits are reached
Tracking complex debugging paths and discovered environment configurations