When is the best time to trigger the memory skill?

It is most effective immediately after a compaction event or when you find yourself needing to recall specific implementation details discussed much earlier in your workflow.

What is memory decay in Claude Code?

Memory decay occurs when the /compact command compresses the conversation to save context window space, which can cause the AI to lose specific details from the beginning of the session.

How does the memory skill prevent repeating errors?

It specifically tracks a file called errors.md which logs previous fix records, allowing Claude to check historical data before attempting a solution that might have already failed.

Will reading large memory files overflow my context window?

No, the skill includes a built-in script that automatically truncates files larger than 15,000 characters, ensuring only the most recent and relevant context is loaded.

Long-term Session Memory

Name: Long-term Session Memory
Author: maxgent-ai

bymaxgent-ai

0•

生产力与工作流

Preserves and retrieves historical context, error records, and code references to prevent information loss during Claude Code session compactions.

The Long-term Session Memory skill solves the problem of 'memory decay' in Claude Code caused by the /compact command. While compaction helps save context window space, it often leads to the loss of crucial early project details and error-fix history. This skill interfaces with a persistent storage system that automatically saves session data to dedicated markdown files, allowing Claude to recall specific technical background, previous debugging steps, and file references even after multiple compactions.

主要功能

01Dedicated error tracking to prevent repeating past debugging mistakes

02Session-specific historical data management

03Categorized memory storage for errors, technical context, and file references

04Automatic file truncation to optimize context window usage

050 GitHub stars

06Persistent context retrieval across session compactions

使用场景

01Restoring critical project context immediately after a /compact command

02Accessing high-level technical requirements discussed at the start of a long development session

03Reviewing error logs to troubleshoot recurring bugs with previously identified solutions

主要功能

01Dedicated error tracking to prevent repeating past debugging mistakes

02Session-specific historical data management

03Categorized memory storage for errors, technical context, and file references

04Automatic file truncation to optimize context window usage

050 GitHub stars

06Persistent context retrieval across session compactions

使用场景

01Restoring critical project context immediately after a /compact command

02Accessing high-level technical requirements discussed at the start of a long development session

03Reviewing error logs to troubleshoot recurring bugs with previously identified solutions