OpenClaw Loom

An OpenClaw plugin that turns your learning into a living graph of nodes, paths, and context captures. Loom blends mastery learning, spaced repetition, and retrieval practice, and indexes everything with QMD for fast search across Markdown and QMD files.

Features

• Learning graph with prerequisites, unlocks, and mastery states
• Spaced repetition scheduling with review ratings
• Next-node recommendations with automatic unlocks
• Context captures to inject real-life situations into your learning map
• Local-first Markdown/QMD library with clean folder structure
• QMD indexing for semantic search and embeddings
• Optional git sync to keep your library in version control

Installation

openclaw plugins install openclaw-plugin-loom

Or from GitHub:

<<<<<<< HEAD
openclaw plugins install github:pepicrft/clawd-plugin-loom
=======
openclaw plugins install github:pepicrft/openclaw-plugin-loom
>>>>>>> 72e4635 (Rename from clawd-plugin-loom to openclaw-plugin-loom)

Configuration

{
  plugins: {
    entries: {
      "openclaw-plugin-loom": {
        enabled: true,
        config: {
          libraryPath: "/Users/you/Loom",
          collectionName: "learn",
          gitRemote: "origin",
          gitBranch: "main",
          gitSync: true,
          gitAutoCommit: true,
          autoInstallQmd: true,
          mask: "**/*.md",
          masteryThreshold: 4,
          srsIntervals: [1, 3, 7, 14, 30, 60, 120, 240]
        }
      }
    }
  }
}

Loom Framework

Folder Convention

• paths/ -> one folder per learning path (Nix, German, etc.)
• nodes/ -> learning nodes grouped by path
• contexts/ -> quick captures from real-life situations
• sessions/ -> auto-logged review sessions
• resources/ -> PDFs, links, or datasets

Notes are plain Markdown files; reference other nodes by id using wikilinks (e.g. [[nix/derivations]]) or normal links.

What Is a Node?

A node is the smallest unit of learning you want to master. It should be narrow, testable, and self-contained.

Good node contents:

• A short explanation in your own words
• A concrete example or mini exercise
• Links to prerequisite or follow-up nodes by id
• If relevant, a "check yourself" prompt (question or task)

Node Frontmatter

Required:

• id
• title
• path
• created
• updated

Recommended:

• summary
• type: concept | practice | project | checkpoint
• status: locked | available | in-progress | mastered | paused
• prerequisites: [node_id]
• unlocks: [node_id]
• familiarity: 0-5
• srs_stage: 0+
• last_reviewed
• next_review
• tags

Example:

---
id: "german/akkusativ"
title: "Akkusativ basics"
summary: "Use Akkusativ with direct objects"
path: "german"
type: "concept"
status: "available"
prerequisites:
  - "german/der-die-das"
unlocks:
  - "german/dative-intro"
familiarity: 1
srs_stage: 0
last_reviewed: null
next_review: null
created: "2026-01-03T12:00:00.000Z"
updated: "2026-01-03T12:00:00.000Z"
tags:
  - "grammar"
---

Body starts here.

Usage

CLI

# Initialize the library structure
openclaw learn init

# Create a path
openclaw learn path "Nix"

# Add a node
openclaw learn node "Nix Store Basics" \
  --path nix \
  --summary "Understand /nix/store layout" \
  --tags nix,foundations

# Capture a contextual situation
openclaw learn capture "German bakery chat" \
  --path german \
  --tags german,speaking \
  "Asked for Schrippen and got confused about plural forms"

# Unlock nodes whose prerequisites are met
openclaw learn unlock

# Suggest next node to study
openclaw learn next --start

# Review a node
openclaw learn review nix/store-basics --rating good

# Query the library
openclaw learn query "nix derivations" --mode query --limit 5

Tools (for agents)

• learn_add_node
• learn_capture
• learn_next
• learn_review
• learn_query

Gateway RPC

• learn.add
• learn.capture
• learn.next
• learn.review
• learn.query
• learn.graph

Git Sync

When git sync is enabled, Loom:

1. Pulls before reading or writing
2. Auto-commits changes (when gitAutoCommit is true)
3. Pushes after writing

QMD Setup

The plugin auto-installs qmd if missing:

• Uses bun install -g https://github.com/tobi/qmd when bun is available
• Falls back to npm install -g https://github.com/tobi/qmd

If you want to manage qmd manually, set autoInstallQmd: false.

SRS Defaults

The default spaced-repetition intervals (days) are:

[1, 3, 7, 14, 30, 60, 120, 240]

You can override them via srsIntervals in the plugin config.

Requirements

• Node 20+
• bun or npm for installing qmd (unless qmd already installed)

Tests

npm test
npm run build

Research Notes

See RESEARCH.md for the learning science sources used to design Loom's framework.

Learning Framework References

The framework is informed by these cross-topic learning principles:

• Distributed practice / spacing effect: Cepeda et al. (2006), Psychological Science. DOI: https://doi.org/10.1111/j.1467-9280.2006.01684.x
• Testing effect / retrieval practice: Roediger & Karpicke (2006), Science. DOI: https://doi.org/10.1126/science.1152408
• Interleaving vs. blocking: Kornell & Bjork (2008), Psychological Science. DOI: https://doi.org/10.1111/j.1467-9280.2008.02127.x
• Mastery learning: Bloom (1968), Learning for Mastery. ERIC: https://files.eric.ed.gov/fulltext/ED053419.pdf
• Deliberate practice: Ericsson, Krampe, & Tesch-Romer (1993), Psychological Review. DOI: https://doi.org/10.1037/0033-295X.100.3.363
• Concept mapping: Novak & Gowin (1984), Learning How to Learn.