How do I trigger a memory search in Cursor?

You can trigger a search by using keywords like 'memory search', 'past decisions', or 'claude-mem' in your prompt, which activates the MCP search tools.

What kind of information is best to record in the memory?

It is most effective for recording non-obvious technical decisions, implementation patterns, specific bug fix histories, and high-level project goals.

Can I use this with both Claude Code and Cursor at the same time?

Yes, the skill is specifically designed for dual-integration. It uses SQLite's WAL mode to handle concurrent access from both tools simultaneously.

Does this skill require an external database or cloud storage?

No, it uses a local SQLite database stored on your machine (~/.claude-mem/claude-mem.db), ensuring your project data remains private and accessible offline.

Cursor Memory Sync

Name: Cursor Memory Sync
Author: Chachamaru127

byChachamaru127

•

122

生産性とワークフロー

Synchronizes persistent project knowledge and implementation decisions between Cursor and Claude Code using a shared MCP memory server.

概要

Cursor Memory Sync creates a unified knowledge layer between the Cursor IDE and Claude Code, ensuring that technical decisions and implementation patterns are never lost between sessions. By integrating the claude-mem MCP server, this skill allows developers to record observations, document the 'why' behind architectural choices, and perform semantic searches over past session histories. It utilizes a shared local SQLite database to provide both tools with a collective memory, enabling the AI to maintain high-level context, recall previous bug fixes, and follow established project conventions without repetitive prompting.

主な機能

Structured observation recording for implementation patterns and bug fixes
122 GitHub stars
Semantic search for retrieving past architectural decisions and rationales
Unified tagging system including source-tracking and entry categorization
Timeline-based context retrieval to track project evolution
Bi-directional memory sharing between Cursor and Claude Code via SQLite

ユースケース

Searching for previous solutions to similar bugs encountered in earlier development phases
Recalling the specific reasoning behind a library choice or architectural pattern during a refactor
Recording 'handoff' notes at the end of a session to guide future autonomous work

概要

主な機能

Structured observation recording for implementation patterns and bug fixes
122 GitHub stars
Semantic search for retrieving past architectural decisions and rationales
Unified tagging system including source-tracking and entry categorization
Timeline-based context retrieval to track project evolution
Bi-directional memory sharing between Cursor and Claude Code via SQLite

ユースケース

Searching for previous solutions to similar bugs encountered in earlier development phases
Recalling the specific reasoning behind a library choice or architectural pattern during a refactor
Recording 'handoff' notes at the end of a session to guide future autonomous work