How does User Memory extract my preferences?

It uses pattern matching via hooks that run after Claude responses to identify phrases regarding your tech stack, tools, and preferred coding style, then saves them to a local JSON profile.

Does it store my data in the cloud?

No, all memory files—including your profile, changelog, and session data—are stored locally on your machine in the ~/.claude/user-memory/ directory.

Can I manually manage or delete my stored data?

Yes, the data is stored in standard JSON format. You can manually edit the files, or use MCP tools like clear_user_profile and remove_preference for granular control.

What is the difference between Minimal and MCP modes?

Minimal mode uses lightweight shell scripts and jq for basic extraction, while MCP mode uses a Node.js server to provide real-time tools, confidence tracking, and advanced session continuity.

How does the preference decay system work?

In MCP mode, preferences have a confidence score that decreases over time if not reinforced. If a preference isn't mentioned for an extended period, it is automatically pruned to keep your profile accurate.

User Memory for Claude Code

Name: User Memory for Claude Code
Author: cameronsjo

bycameronsjo

•

생산성 및 워크플로

Persists user preferences, tech stacks, and session context across Claude Code interactions to provide a personalized, long-term development profile.

소개

The User Memory skill transforms Claude Code into a context-aware assistant by building a persistent profile of your development habits and technical preferences. It automatically extracts information such as your preferred tech stack, editor, and communication tone from your conversations without requiring manual input. By utilizing hooks and an optional MCP server, it manages session continuity, tracks project decisions, and implements a confidence-based decay system to ensure your profile remains relevant. This eliminates the need to repeat setup instructions and allows Claude to maintain a consistent 'memory' of how you work across different sessions and projects.

주요 기능

Automatic preference extraction from conversation patterns with zero latency.
Dual-mode implementation offering lightweight shell scripts or a full TypeScript MCP server.
Session continuity tracking for task progress, technical decisions, and cross-session summaries.
Confidence-based decay system that automatically phases out stale or outdated preferences.
Detailed changelog and audit trail for all profile modifications and extractions.
1 GitHub stars

사용 사례

Maintaining consistent coding standards and editor preferences across multiple disjointed sessions.
Resuming complex development tasks where specific architectural decisions need to be remembered.
Personalizing AI response tone and detail level based on your specific professional role.

소개

주요 기능

Automatic preference extraction from conversation patterns with zero latency.
Dual-mode implementation offering lightweight shell scripts or a full TypeScript MCP server.
Session continuity tracking for task progress, technical decisions, and cross-session summaries.
Confidence-based decay system that automatically phases out stale or outdated preferences.
Detailed changelog and audit trail for all profile modifications and extractions.
1 GitHub stars

사용 사례

Maintaining consistent coding standards and editor preferences across multiple disjointed sessions.
Resuming complex development tasks where specific architectural decisions need to be remembered.
Personalizing AI response tone and detail level based on your specific professional role.