Skillget

memesh-llm-memory

mcp
Guvenlik Denetimi
Uyari
Health Uyari
  • License — License: MIT
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Low visibility — Only 6 GitHub stars
Code Gecti
  • Code scan — Scanned 12 files during light audit, no dangerous patterns found
Permissions Gecti
  • Permissions — No dangerous permissions requested
Purpose
This plugin provides persistent memory for Claude Code and other compatible AI coding assistants. It uses a local knowledge graph to store and recall context, decisions, and patterns across different coding sessions.

Security Assessment
Overall risk: Low. The tool stores all data locally in an SQLite database located in the user's home directory (`~/.memesh/knowledge-graph.db`), meaning no sensitive data leaves your machine. The automated code scan across 12 files found no dangerous patterns, hardcoded secrets, or requests for elevated system permissions. It operates offline with a strict zero-network-request architecture. While it does utilize background hooks to record git commits and auto-recall memories, these are standard local operations directly related to its intended purpose and pose no security threat.

Quality Assessment
The project is actively maintained, with its most recent code push occurring today. It is distributed under the standard MIT license, which is excellent for open-source adoption. However, community visibility is currently very low, with only 6 GitHub stars. Being a newer or niche project, it lacks the extensive peer review and battle-testing that more popular tools receive. The provided documentation is highly detailed, clearly explaining its setup, architecture, and comparisons to native alternatives.

Verdict
Safe to use.
SUMMARY

Persistent memory plugin for Claude Code with knowledge graph, semantic search, hooks, and session recall.

README.md

MeMesh Plugin

Minimal persistent memory plugin for Claude Code. Remembers decisions, patterns, and context across sessions.

npm version
License
Node
MCP

Why MeMesh?

Anthropic ships a built-in Memory Tool in the Messages API. MeMesh solves a different problem:

MeMesh Anthropic Memory Tool
Data location Local SQLite on your machine Anthropic cloud
Data model Knowledge graph (entities + relations + tags) Key-value store
Search FTS5 full-text search Exact key lookup
Visualization Interactive D3.js dashboard None
Integration Claude Code plugin with auto-hooks Messages API parameter
Privacy 100% offline, zero data leaves your machine Data sent to Anthropic
Extensibility Open source, hackable SQLite Closed API

Choose MeMesh when: you want local-first, structured knowledge that survives across projects, with full-text search and visual exploration.

Choose Anthropic Memory Tool when: you want zero-setup cloud memory tightly integrated with the Messages API.

Installation

npm install -g @pcircle/memesh

What it does

MeMesh gives Claude Code persistent memory through 3 MCP tools, 2 hooks, and a CLI dashboard:

MCP Tools

Tool Description
remember Store knowledge — entities with observations, relations, and tags
recall Search stored knowledge via FTS5 full-text search with optional tag filtering
forget Archive knowledge (soft-delete) or remove specific observations

Hooks

Hook Event What
Session Start SessionStart Auto-recalls project-specific + recent global memories
Post Commit PostToolUse (Bash) Records git commits as knowledge entities

CLI

Command Description
memesh-view Generate and open an interactive HTML dashboard with bundled local D3 
memesh-view

MeMesh Dashboard

How it works

  • Storage: SQLite database at ~/.memesh/knowledge-graph.db
  • Search: FTS5 full-text search (no vector embeddings)
  • Isolation: Tag-based project filtering (project:<name>)
  • Upserts: Reusing an entity name appends observations, preserves the original type, and dedupes tags
  • Dashboard: memesh-view bundles D3 locally, so the generated HTML works offline
  • Schema: entities, observations, relations, tags + FTS5 virtual table
  • Validation: All tool inputs validated with Zod schemas
  • Knowledge Evolution: forget archives rather than deletes — old memories are preserved but hidden. Use supersedes relations to replace old designs with new ones.

Architecture

src/
├── cli/
│   └── view.ts           # HTML dashboard generator (D3.js graph + stats)
├── db.ts                 # SQLite database (open/close/migrate, FTS5)
├── knowledge-graph.ts    # Entity CRUD, relations, FTS5 search
├── index.ts              # Package exports
└── mcp/
    ├── server.ts         # MCP server entry point (stdio transport)
    └── tools.ts          # 3 tool handlers + Zod validation

scripts/hooks/
├── session-start.js      # Auto-recall on session start
└── post-commit.js        # Git commit tracking

Dependencies (3): better-sqlite3, @modelcontextprotocol/sdk, zod

Development

npm install
npm run build       # TypeScript compile + bundle assets + set executables
npm test -- --run   # Run the vitest suite once
npm run typecheck   # tsc --noEmit
npm run test:packaged   # Pack/install smoke test for the published artifact

Project Policies

License

MIT

Yorumlar (0)

Sonuc bulunamadi