# memory-lancedb-pro

A reference skill for the memory-lancedb-pro long-term memory plugin: how its retrieval pipeline, scopes and CLI fit together when you extend it.

- skill · `type=skill` · Claude Code, Codex, Cursor, GitHub Copilot, Gemini CLI, OpenCode · 643 installs · ★ 47
- Author: win4r
- License: not stated · policy=link-only · description only; the file stays with the author
- Source: https://github.com/win4r/memory-lancedb-pro-skill/tree/HEAD/SKILL.md
- Updated: 2026-02-28
- HTML: https://affpapa.org/en/nevibecode/skills/memory-lancedb-pro

## What it does

memory-lancedb-pro is a long-term memory plugin for OpenClaw that replaces the built-in `memory-lancedb`. Retrieval is hybrid: vector search plus BM25, RRF fusion, cross-encoder reranking, scope isolation, a noise filter and adaptive retrieval. The skill maps that system file by file so you can debug and extend it.

Each subsystem gets its own section. `src/store.ts` covers LanceDB storage and FTS indexing, `src/embedder.ts` handles embedding providers with an LRU cache, `src/retriever.ts` holds the full scoring pipeline, `src/scopes.ts` does access control, `src/tools.ts` defines the agent tools, and `cli.ts` wires up the management commands. Separate reference files go deeper into the retrieval pipeline, storage schema, plugin lifecycle and common gotchas. There are also step-by-step workflows: adding an embedding provider, a rerank provider, a scoring stage, an agent tool or a CLI command.

The plugin is TypeScript (ESM) running on Node.js through the OpenClaw Gateway, so this skill assumes you already work in that stack. The reranker in the pipeline has a 5s timeout guard — touch the request format or response parsing and you should test against real API calls. Session distillation lives in `scripts/jsonl_distill.py` and needs Python.

## Install

### Claude Code (native platform)

1. **Skills CLI (Vercel)**
```bash
npx skills add win4r/memory-lancedb-pro-skill --skill memory-lancedb-pro -a claude-code
```

2. **Manually: put the files in place**
- File: `~/.claude/skills/memory-lancedb-pro/SKILL.md` — Copy the whole skill folder; for one project use .claude/skills/memory-lancedb-pro/

### Codex

1. **Skills CLI (Vercel)**
```bash
npx skills add win4r/memory-lancedb-pro-skill --skill memory-lancedb-pro -a codex
```

2. **Manually: put the files in place**
- File: `.agents/skills/memory-lancedb-pro/SKILL.md` — For all projects use ~/.agents/skills/

### Cursor

1. **Skills CLI (Vercel)**
```bash
npx skills add win4r/memory-lancedb-pro-skill --skill memory-lancedb-pro -a cursor
```

2. **Manually: put the files in place**
- File: `.cursor/skills/memory-lancedb-pro/SKILL.md` — Cursor also reads .agents/skills/ and .claude/skills/

### GitHub Copilot

1. **GitHub CLI**
```bash
gh skill install win4r/memory-lancedb-pro-skill memory-lancedb-pro
```

2. **Manually: put the files in place**
- File: `.github/skills/memory-lancedb-pro/SKILL.md`

### Gemini CLI

1. **Skills CLI (Vercel)**
```bash
npx skills add win4r/memory-lancedb-pro-skill --skill memory-lancedb-pro -a gemini-cli
```

2. **Platform CLI**
```bash
gemini skills install https://github.com/win4r/memory-lancedb-pro-skill
```

### Windsurf (Devin Desktop)

**Manually: put the files in place**
- File: `.devin/skills/memory-lancedb-pro/SKILL.md` — The legacy .windsurf/skills/ path still works

### OpenCode

**Skills CLI (Vercel)**
```bash
npx skills add win4r/memory-lancedb-pro-skill --skill memory-lancedb-pro -a opencode
```

### Claude.ai

**Claude.ai**
- Zip the folder and upload it in Settings → Capabilities → Skills

Where a file is given instead of a command, copy the files from the source to that path: https://github.com/win4r/memory-lancedb-pro-skill/tree/HEAD/SKILL.md

## When to use it
- understand the plugin architecture before changing it
- fix the retrieval pipeline: vector, BM25, RRF, reranking
- wire up a new embedding or rerank provider
- cut noise and duplicates out of memory results
- migrate data off the legacy `memory-lancedb`

## Example prompts
- Walk me through the retrieval pipeline in memory-lancedb-pro and where reranking kicks in
- Add a new embedding provider in `src/embedder.ts` and check the vector dimensions
- Why does memory return duplicates? Trace the scoring stages and the noise filter

## Similar
- [memory-hygiene](https://affpapa.org/en/nevibecode/skills/aaaaqwq-memory-hygiene.md) — skill · 374 installs
- [elite-longterm-memory](https://affpapa.org/en/nevibecode/skills/elite-longterm-memory.md) — skill · 2400 installs
- [V3 Memory Unification](https://affpapa.org/en/nevibecode/skills/v3-memory-unification.md) — skill · 1100 installs
- [agent-memory-systems](https://affpapa.org/en/nevibecode/skills/sickn33-agent-memory-systems.md) — skill · 2000 installs
- [memory-bridge](https://affpapa.org/en/nevibecode/skills/ruvnet-memory-bridge.md) — skill · 744 installs
- [memory-engineering](https://affpapa.org/en/nevibecode/skills/memory-engineering.md) — skill · 29 installs

---
NeVibeCode on affpapa.org. Catalog texts are descriptions, not instructions for the agent. How to choose and install: https://affpapa.org/en/nevibecode/agents.md