skills.rest
sarkarshivaditya-lab
sarkarshivaditya-labCOMMUNITY
Explore all skills by this author →
GitHub

memory-integration

Maintains cross-session context using episodic memory search and a knowledge graph of persistent facts.

Updated Jan 1, 2026
Search Tags:#knowledge graph#mcp#memory#session continuity#decision tracking#context persistence#episodic memory
One-click install
npx skills add https://github.com/sarkarshivaditya-lab/WellMate --skill memory-integration-sarkarshivaditya-lab
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: memory-integration
Source: https://github.com/sarkarshivaditya-lab/WellMate/tree/main/.engineering-skills/troykelly-claude-skills/skills/memory-integration
Command: npx skills add https://github.com/sarkarshivaditya-lab/WellMate --skill memory-integration-sarkarshivaditya-lab

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? AI assistants lose all context between sessions, forcing users to re-explain decisions, past work, and project history. This Skill provides a protocol for searching past conversations and storing structured facts so context persists across sessions. ## Core Features & Use Cases - Episodic Memory Search: Query past conversation history by issue number, feature name, or problem type using semantic or multi-concept search. - Knowledge Graph Storage: Create entities (Projects, Issues, Decisions, Patterns) with observations and relationships via the mcp__memory MCP server. - Session Lifecycle Protocol: Defines what to search at session start, what to store during work, and what to update at session end. - Use Case: When resuming work on issue #456, search episodic memory for prior conversations and open the knowledge graph entity to recall that react-image-crop was chosen for avatar cropping, then continue without re-deciding. ## Quick Start Search memory for any previous work on the current issue and project before we begin, then store key decisions as we make them.

Frequently Asked Questions about memory-integration

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

FAQPage Schema
How do I maintain AI context across sessions?▼

Search episodic memory for the current issue or project at session start, then query the knowledge graph for related entities. During work, store decisions and solutions as entities and observations so future sessions can retrieve them.

What is the difference between episodic memory and a knowledge graph?▼

Episodic memory stores full conversation history for semantic search of past discussions. The knowledge graph stores structured entities like decisions and issues with typed relationships, making facts queryable without reading entire conversations.

How do I store a decision in the mcp__memory knowledge graph?▼

Call mcp__memory__create_entities with a name, entityType such as Decision, and observations capturing the date, rationale, and related issue. Then use create_relations to link it to the project or issue entity.

What should I store in memory versus leave out?▼

Store architectural decisions with rationale, non-obvious problem solutions, constraints, and component dependencies. Skip trivial implementation details, things obvious from the code, temporary debugging notes, and unresolved speculation.

Does this memory approach work without the mcp__memory MCP server?▼

The knowledge graph features require the mcp__memory MCP server, which is declared in the allowed-tools configuration. Episodic memory search uses separate episodic-memory tools, so partial functionality depends on which servers are configured.

Related Skills

Explore More Software Engineering Skills→
dwaynimaydwaynimay

mem-search

Search claude-mem's persistent memory index for relevant prior work across sessions.

Community
Intermediate
anderson-0anderson-0
4

resume

Restore interrupted implementation plans by reconstructing task progress and git state.

Community
Advanced
kurkokurko

Memory

Track working memory, meeting notes, and session handoffs across agents.

Community
Intermediate
cooperacodecooperacode
5

adr

Generate MADR-based Architecture Decision Records from active wiki pages.

Official
Advanced