Instructional Design Skills Kit

11 OpenClaw skills for Instructional Designers covering the full ADDIE lifecycle — from needs analysis to SCORM delivery, plus expert-level learning science advisory and design. Use the agent/ workspace for multi-agent deployment.

Status: v0.1.0

Architecture

Agent:              agent/ (OpenClaw agent workspace — AGENTS.md, SOUL.md, IDENTITY.md)
Lobsters:           id-full-course | id-quick-lesson | id-deploy | id-review-cycle
Layer 0 - Orchestrator: id-orchestrator (lobster-based course build coordination)
Layer 1 - Core:     id-core (pipeline management, shared knowledge)
Layer 2 - Domain:   id-analyst | id-creator | id-builder | id-operator
Layer 3 - Integration: id-lms | id-google | id-notion
Expert:             id-expert (learning science advisory + hands-on design)
Setup:              id-setup (onboarding wizard)

Skills

Lobster Workflows

The kit includes 4 lobster workflows that chain id-* skills into automated pipelines:

Workflow files live in lobsters/. For dynamic course builds, task-decomposer.py --format lobster generates a course-specific workflow with the correct number of lessons.

Agent (Multi-Agent Deployment)

The agent/ directory contains OpenClaw agent workspace files for running the kit as a dedicated agent with its own identity, memory, and sessions.

Quick Setup

# Add a new agent
openclaw agents add id-agent --workspace ~/.openclaw/workspace-id-agent

# Copy workspace files
cp agent/AGENTS.md agent/SOUL.md agent/IDENTITY.md agent/TOOLS.md agent/BOOTSTRAP.md \
   ~/.openclaw/workspace-id-agent/

# Install skills
for skill in skills/id-*; do
  ln -sf "$(pwd)/$skill" ~/.openclaw/skills/$(basename $skill)
done

Workspace Files

Multi-Agent Routing

Run alongside your main agent — route by channel, sender, or group:

{
  "agents": {
    "list": [
      { "id": "main", "default": true, "workspace": "~/.openclaw/workspace" },
      { "id": "id-agent", "workspace": "~/.openclaw/workspace-id-agent" }
    ]
  },
  "bindings": [
    { "agentId": "id-agent", "match": { "channel": "telegram", "accountId": "id-bot" } }
  ]
}

See agent/README.md for full setup guide and OpenClaw multi-agent docs for routing details.

Requirements

• Node.js 22+
• Python 3.9+
• Optional Python packages (see Install Python dependencies)

Deployment

Option A: OpenClaw Runtime (Recommended)

1. Install OpenClaw

macOS / Linux:

curl -fsSL https://openclaw.ai/install.sh | bash

Windows (PowerShell):

iwr -useb https://openclaw.ai/install.ps1 | iex

Verify installation:

node --version    # Must be 22+
openclaw gateway status

2. Run onboarding

openclaw onboard --install-daemon

This configures auth, gateway, and channels. Dashboard available at http://127.0.0.1:18789/.

3. Install skills

Clone the repo and symlink skills into OpenClaw:

git clone https://github.com/hoadh/instructional-design-agentic-kit.git
cd instructional-design-agentic-kit

# Symlink all skills
for skill in skills/id-*; do
  ln -sf "$(pwd)/$skill" ~/.openclaw/skills/$(basename $skill)
done

Alternatively, copy skills directly:

cp -r skills/id-* ~/.openclaw/skills/

Or add this repo's skills/ folder as an extra skills directory in ~/.openclaw/openclaw.json:

{
  "skills": {
    "load": {
      "extraDirs": ["/path/to/instructional-design-agentic-kit/skills"]
    }
  }
}

4. Install Python dependencies

pip install httpx jinja2 pandas lxml google-genai google-auth google-auth-oauthlib google-api-python-client

5. Configure integrations

Run the setup wizard inside an OpenClaw session:

/id-setup

Or configure manually by setting environment variables in ~/.openclaw/openclaw.json:

{
  "skills": {
    "entries": {
      "id-lms": {
        "enabled": true,
        "env": {
          "ID_LMS_PLATFORM": "moodle",
          "ID_LMS_URL": "https://your-lms.example.com",
          "ID_LMS_TOKEN": "your-api-token"
        }
      },
      "id-google": {
        "enabled": true,
        "env": {
          "ID_GOOGLE_CREDENTIALS": "~/.openclaw/credentials/google-oauth.json",
          "ID_GOOGLE_TOKEN": "~/.openclaw/credentials/google-token.json",
          "ID_GOOGLE_FOLDER_ID": "optional-drive-folder-id"
        }
      },
      "id-notion": {
        "enabled": true,
        "env": {
          "ID_NOTION_TOKEN": "ntn_your-token",
          "ID_NOTION_DATABASE_ID": "your-database-id"
        }
      }
    }
  }
}

6. Verify

/id-setup check

This runs the health check: Python version, packages, skill installation, integrations.

7. Create your first project

Option A: Full orchestration (recommended)

/id-build "Build a 5-lesson AI Ethics course for product managers, deploy to Moodle" --mode auto

The orchestrator decomposes this into a task graph and coordinates all skills automatically, pausing at ADDIE phase boundaries for your review.

Option B: Direct lobster workflow

openclaw lobster run id-quick-lesson --topic "AI Ethics Module 1" --course ai-ethics

Option C: Manual step-by-step

/id-pipeline new "AI Ethics for Product Managers" --lang en
/id-analyze
/id-lesson
/id-scorm
/id-review
/id-launch

8. Connect to Telegram (optional)

Build courses from your phone or team group chat.

1. Create a bot via @BotFather and save the token.

2. Add to ~/.openclaw/openclaw.json:

{
  "channels": {
    "telegram": {
      "enabled": true,
      "botToken": "YOUR_BOT_TOKEN",
      "dmPolicy": "pairing",
      "customCommands": [
        { "command": "build", "description": "Build a course" },
        { "command": "status", "description": "Check progress" }
      ]
    }
  }
}

3. Approve your user: openclaw pairing approve

4. Chat with your bot: "Build me a 5-lesson compliance course for nurses, deploy to Moodle"

See docs/examples.md for detailed Telegram usage examples.

Option B: Docker

Standalone container

docker build -t id-skills-kit .
docker run -it \
  -p 18789:18789 \
  -v id-skills-workspace:/root/.openclaw/workspace \
  -v id-skills-memory:/app/memory \
  -e ID_LMS_PLATFORM=moodle \
  -e ID_LMS_URL=https://your-lms.example.com \
  -e ID_LMS_TOKEN=your-api-token \
  -e ID_GOOGLE_CREDENTIALS=/run/secrets/google-creds.json \
  -e ID_NOTION_TOKEN=secret_your-token \
  -e ID_NOTION_DATABASE_ID=your-database-id \
  id-skills-kit

Key volume mounts:

• /root/.openclaw/workspace — OpenClaw memory and state (persists between restarts)
• /app/memory — Project pipeline data (course content, builds, configs)

Docker Compose

Create docker-compose.yml:

services:
  id-skills-kit:
    build: .
    ports:
      - "18789:18789"
    volumes:
      - openclaw-workspace:/root/.openclaw/workspace
      - project-memory:/app/memory
      - ${ID_GOOGLE_CREDENTIALS:-/dev/null}:/run/secrets/google-creds.json:ro
    environment:
      - OPENCLAW_HOME=/root/.openclaw
      # LMS
      - ID_LMS_PLATFORM=${ID_LMS_PLATFORM:-}
      - ID_LMS_URL=${ID_LMS_URL:-}
      - ID_LMS_TOKEN=${ID_LMS_TOKEN:-}
      # Google Workspace
      - ID_GOOGLE_CREDENTIALS=/run/secrets/google-creds.json
      - ID_GOOGLE_FOLDER_ID=${ID_GOOGLE_FOLDER_ID:-}
      # Notion
      - ID_NOTION_TOKEN=${ID_NOTION_TOKEN:-}
      - ID_NOTION_DATABASE_ID=${ID_NOTION_DATABASE_ID:-}
    restart: unless-stopped

volumes:
  openclaw-workspace:
  project-memory:

Create .env file (never commit this):

# LMS Configuration
ID_LMS_PLATFORM=moodle
ID_LMS_URL=https://your-lms.example.com
ID_LMS_TOKEN=your-api-token

# Google Workspace
ID_GOOGLE_CREDENTIALS=~/.openclaw/credentials/google-oauth.json
ID_GOOGLE_TOKEN=~/.openclaw/credentials/google-token.json
ID_GOOGLE_FOLDER_ID=

# Notion
ID_NOTION_TOKEN=ntn_your-token
ID_NOTION_DATABASE_ID=your-database-id

Run:

docker compose up -d
docker compose exec id-skills-kit openclaw dashboard

Dockerfile

Create Dockerfile:

FROM node:22-slim

# Install Python and pip
RUN apt-get update && \
    apt-get install -y --no-install-recommends python3 python3-pip curl && \
    rm -rf /var/lib/apt/lists/*

# Install OpenClaw
RUN curl -fsSL https://openclaw.ai/install.sh | bash

# Install Python dependencies
RUN pip3 install --break-system-packages httpx jinja2 pandas lxml google-genai google-auth google-auth-oauthlib google-api-python-client

# Copy skills into the managed skills directory
WORKDIR /app
COPY skills/ /root/.openclaw/skills/

# Configure OpenClaw to use /app/memory for project data
ENV OPENCLAW_HOME=/root/.openclaw

EXPOSE 18789

CMD ["openclaw", "onboard", "--install-daemon"]

Build and run:

docker build -t id-skills-kit .
docker compose up -d

Google Workspace Setup

Two authentication methods are supported:

Method A: OAuth2 Desktop (Recommended)

Best for Google Workspace orgs that enforce iam.disableServiceAccountKeyCreation policy (blocks service account keys).

1. Go to Google Cloud Console → APIs & Services → Credentials
2. Create an OAuth Client ID → Application type: Desktop app
3. Enable APIs: Google Docs, Google Sheets, Google Forms, Google Drive
4. Download the client credentials JSON → save as ~/.openclaw/credentials/google-oauth.json
5. Run the OAuth flow to obtain access + refresh tokens:

   python3 skills/id-google/scripts/google-oauth-flow.py
   

   This saves tokens to ~/.openclaw/credentials/google-token.json
6. Set environment variables:
   • ID_GOOGLE_CREDENTIALS → path to google-oauth.json
   • ID_GOOGLE_TOKEN → path to google-token.json

Method B: Service Account (Original)

⚠️ May be blocked by org policy (iam.disableServiceAccountKeyCreation). Use Method A if this fails.

1. Create a service account in Google Cloud Console
2. Download the JSON key → set ID_GOOGLE_CREDENTIALS to its path
3. Share target Drive folders / Docs with the service account email

Notion Setup

1. Create an internal integration
2. Share target pages/databases with the integration
3. Set ID_NOTION_TOKEN to the integration token

Note: Notion now issues tokens with ntn_ prefix (new format). Both secret_ and ntn_ prefixes work.

Media Generation

Image Generation

Two methods are available:

API Method

Get an API key from Google AI Studio (requires billing enabled on GCP project for image generation).

export GEMINI_API_KEY=your-key
python3 skills/id-creator/scripts/image-generator-api.py \
  --prompts-file memory/projects/<course>/media/media-prompts.md \
  --output-dir memory/projects/<course>/media/generated/ \
  --delay 3

Browser Method

• Requires: OpenClaw node with Chrome + Browser Relay extension
• Uses Google One Ultra credits (completely free)
• See skills/id-creator/scripts/image-generator-browser.md

Prompt Parser

Parse media-prompts.md into structured prompts:

python3 skills/id-creator/scripts/prompt-parser.py \
  --input media-prompts.md \
  --style dalle \
  --output prompts.json

Drive Upload

Upload files to Google Drive with automatic deduplication (safe to retry):

python3 skills/id-google/scripts/drive-uploader.py \
  --parent-folder-id "FOLDER_ID" \
  --folder-name "Phase 1" \
  --files needs-analysis.md:"Needs Analysis" personas.md:"Personas"

• Prevents duplicate folders/files on retry
• Requires ~/.openclaw/credentials/google-token.json (see OAuth2 Desktop setup)

Configuration Reference

Scripts Quick Reference

ID Expert — Learning Science Advisory & Design

Standalone dual-mode agent: advisory (review, consult, improve) and hands-on (design, create). Works with any file — no pipeline dependency.

Advisory Mode

# Rubric-scored quality review of any artifact
/id-review content/lessons/module-1-intro.md --focus engagement

# Before/after improvement suggestions
/id-improve content/quizzes/quiz-03.md --aspect assessment

# Evidence-based consultation
/id-consult "How should I assess soft skills in an async online course?"

# Theory/framework recommendation for a specific context
/id-theory "Onboarding program for 200 remote sales reps, 2-week timeline"

Hands-on Mode

# End-to-end learning experience design
/id-design "Data Privacy Fundamentals" --approach action-mapping

# Learner journey map with engagement design
/id-lxd "New managers conduct effective 1-on-1 meetings within 30 days"

# Job aid / performance support design
/id-job-aid "Troubleshooting customer billing disputes"

CLI Scripts

# Review an artifact file directly
python3 skills/id-expert/scripts/expert-reviewer.py --file lesson.md --focus engagement

# Generate a design document
python3 skills/id-expert/scripts/experience-designer.py --topic "AI Ethics" --approach addie --output design.md

Knowledge Base

Covers: Cognitive Load Theory (Sweller), Multimedia Learning (Mayer's 12 principles), Adult Learning (Knowles' andragogy), ARCS Motivation (Keller), evidence-based learning science (spacing, retrieval, interleaving, elaboration, dual coding), Learning Experience Design (Kolb + 5 Moments of Need), quality rubrics (1–4 scoring), and a theory selection decision matrix.

License

MIT