Nemp Memory
Privacy-first memory plugin for Claude Code. 22 command definitions, cortex vitality scoring, gold test suite, 100% local plain JSON — zero runtime deps.
Choose how to use this skill
You do not need every option. Choose the path your AI client supports. The stable page stays the same; versioned files are immutable.
1. Native installer
This listing has no registered native installer command. Use the complete package or source fallback below, depending on what your client supports.
Do not guess an installer command or replace an existing version without reviewing the diff.
2. Complete package recommended
Download the ZIP when available. It includes SKILL.md plus the references, security notes and version metadata.
No complete ProSkills package is published for this listing yet.3. Prompt-only
Copy the prompt above when the agent can read the stable page or when you want to adopt the workflow without installing a skill.
Need only the instruction file?
Download SKILL.md only if your client requires a single file. The complete ZIP is safer for a full installation because it preserves the references and release context.
No path installs or executes anything by itself. Your agent still needs access to the project files. Before updating, compare the installed version and review the diff.
// RATINGS
Not yet listed on ClawHub or SkillsMP
// README
The Problem
Claude Code forgets everything between sessions. You waste 15-20 minutes every day re-explaining:
- Your tech stack
- Architecture decisions
- Project patterns
- What you worked on yesterday
Other memory plugins exist, but they're complicated:
- Require MCP servers, SQLite databases, or Ollama
- Need cloud accounts and API keys
- Send your code context to their servers
- Need 10+ steps to set up
- Come with web UIs you'll never use
- Store data in formats you can't easily read
Why Nemp Is Different
Nemp is stupidly simple:
/plugin marketplace add https://github.com/SukinShetty/Nemp-memory
/plugin install nemp
# Done.
Zero dependencies. No cloud. No API keys. Plain JSON files. Just works.
Cross-platform: Works on Claude Code, OpenClaw, and any AgentSkills-compatible platform. One memory store, multiple agents.
| Feature | Other Plugins | Nemp |
|---|---|---|
| Setup | 10+ steps | 2 commands |
| Dependencies | SQLite, Ollama, web servers | None |
| Cloud Required | Often yes | No |
| API Key Required | Often yes | No |
| Data Storage | Binary databases, cloud | Plain JSON |
| Privacy | Data may leave your machine | 100% Local |
| Auto-detect Stack | No | Yes |
| Proactive Suggestions | No | Yes |
| Auto-sync to CLAUDE.md | No | Yes |
| Two-way CLAUDE.md sync | No | Yes |
| Conflict detection | No | Yes |
| Works Offline | Sometimes | Always |
| Cost | Free to $19/month | Free Forever |
What's New in v0.2.0
| Feature | Description |
|---|---|
| Agent ID Tracking | Every memory records which agent wrote it — main, nemp-init, backend, etc. |
| Access Logs | All read/write/delete operations logged to .nemp/access.log with timestamps and agent names |
/nemp:log Command | View the full audit trail. Filter by agent, tail recent entries, or clear the log |
| Token Compression | /nemp:save compresses values before storing — ~70% smaller, preserves all technical terms |
/nemp:init Optimized | Single bash scan, reads only package.json, writes all memories in one operation |
| MEMORY.md Index | Auto-generated index at .nemp/MEMORY.md with agent attribution and quick overview |
6 Features That Set Nemp Apart
1️⃣ Auto-Init: One Command Learns Everything
Unique to Nemp: Auto-detects your entire stack in one command.
/nemp:init
That's it. Nemp scans your project and automatically detects:
- Framework (Next.js, React, Vue, etc.)
- Language & config (TypeScript, strict mode)
- Database & ORM (Prisma, Drizzle, MongoDB)
- Auth solution (NextAuth, Clerk, Supabase)
- Styling (Tailwind, styled-components)
- Package manager (npm, yarn, pnpm, bun)
Example output:
Scanning your project...
I found:
Framework: Next.js 14 (App Router detected)
Language: TypeScript (strict mode enabled)
Database: PostgreSQL via Prisma
Auth: NextAuth.js
Styling: Tailwind CSS
Package Manager: npm
Saved 6 memories in 2 seconds
Claude now knows your stack forever.
Why this matters:
- New team members: Zero onboarding
- Context switching: Instant project recall
- No manual documentation needed
2️⃣ Smart Context: Find Memories Instantly
Unique to Nemp: Semantic search that understands what you're looking for.
/nemp:context auth
Nemp doesn't just search for "auth" — it expands to:
- authentication, login, session, jwt, oauth, nextauth, clerk, token, passport, credentials...
Example output:
FOUND 3 MEMORIES MATCHING "auth"
auth-provider [KEY MATCH]
NextAuth.js with JWT strategy
auth-tokens [KEY + VALUE MATCH]
15min access tokens, 7day refresh
auth-middleware [KEY MATCH]
Protects all /api routes except /auth/*
Quick actions:
/nemp:recall auth-provider # View details
/nemp:context database # Search database
Why this matters:
- No scrolling through old chats
- Context appears in < 1 second
- Finds related memories automatically
3️⃣ Memory Suggestions: AI Suggests What to Save
Unique to Nemp: Nemp watches your work and proactively suggests memories.
/nemp:suggest
Example output:
NEMP MEMORY SUGGESTIONS
Based on your recent activity patterns:
#1 auth-approach PRIORITY: HIGH
DRAFTED FOR YOU:
Authentication: JWT tokens, 15min access,
7day refresh. Files: login.ts, session.ts,
middleware.ts in auth/ directory
WHY SUGGESTED:
You edited 3 auth files 7+ times in 30 minutes.
This pattern is worth remembering.
[1] Save [E] Edit [S] Skip
What it detects:
- Files edited frequently (3+ times)
- New packages installed (npm install)
- Directory patterns (auth/, api/, components/)
- Command patterns (test before commit)
- Time-based focus (30+ min sessions)
Why this matters:
- Nemp drafts memories FOR you
- Zero cognitive load
- Captures patterns you'd miss
4️⃣ CLAUDE.md Auto-Sync: Set It and Forget It
Unique to Nemp: Your CLAUDE.md updates itself every time you save a memory.
/nemp:auto-sync on
That's it. From this point:
/nemp:saveadds a memory → CLAUDE.md updates automatically/nemp:initdetects your stack → CLAUDE.md updates automatically/nemp:forgetremoves a memory → CLAUDE.md updates automatically
You never manually edit the project context section of CLAUDE.md again.
What it looks like:
/nemp:save testing "test-gated success with pass@k evaluation"
✓ Memory saved: testing
Value: "test-gated success with pass@k evaluation"
Location: .nemp/memories.json (project)
Total memories: 12
✓ CLAUDE.md synced
Your CLAUDE.md now has two parts:
- Top: Your rules (written by you, never touched by Nemp)
- Bottom: Project context (auto-generated by Nemp, always current)
Why this matters:
- Two commands on day 1. Zero commands after that.
- CLAUDE.md stays accurate without you thinking about it
- New session? Claude already knows your full project context
5️⃣ Two-Way Sync: CLAUDE.md ↔ Nemp
Unique to Nemp: Reads FROM your CLAUDE.md and writes BACK to it.
/nemp:sync
What it does:
Direction 1 — Imports your manually written CLAUDE.md notes into Nemp. Now those notes become searchable with /nemp:context.
Direction 2 — Checks your actual project files (package.json, config files) against what CLAUDE.md says. If something doesn't match, Nemp catches it.
Example:
/nemp:sync
Imported from CL
// HOW IT'S BUILT
KEY FILES