Canvas API Connection Skill

Complete guide to connecting and using Canvas LMS API for Whitecliffe courses.

Quick Reference

Current Courses (2025)

Connection Methods

Method 1: Canvas MCP (Recommended)

The Canvas MCP server is already configured and provides 12 tools:

Course Tools:

• get_courses - List all enrolled courses
• get_modules - Get course modules

Assignment Tools:

• get_assignments - List assignments for a course
• get_upcoming_assignments - Next 7 days deadlines

Grade Tools:

• get_grades - Current grades for a course
• get_submissions - Submission details

Discussion Tools:

• get_discussion_topics - List discussions
• get_announcements - Course announcements

Quiz Tools:

• get_quizzes - List quizzes

Example Usage

# Get all courses
mcp__canvas-mcp__get_courses()

# Get assignments for IT8107
mcp__canvas-mcp__get_assignments(course_id=2366)

# Check grades
mcp__canvas-mcp__get_grades(course_id=2366)

# Get upcoming deadlines
mcp__canvas-mcp__get_upcoming_assignments()

Method 2: Direct API

For custom integrations or when MCP is unavailable:

Base URL: https://learn.mywhitecliffe.com/api/v1

Authentication:

# Store token securely (never commit!)
export CANVAS_TOKEN="your-token-here"

# API request
curl -H "Authorization: Bearer $CANVAS_TOKEN" \
     "https://learn.mywhitecliffe.com/api/v1/courses"

Common Endpoints:

Method 3: Python SDK

from canvasapi import Canvas

API_URL = "https://learn.mywhitecliffe.com"
API_KEY = os.environ.get("CANVAS_TOKEN")

canvas = Canvas(API_URL, API_KEY)
user = canvas.get_current_user()
courses = user.get_courses(enrollment_state='active')

Setup Instructions

Adding Canvas MCP to Claude

Already configured in mcp-config.json:

{
  "mcpServers": {
    "canvas-mcp": {
      "type": "sse",
      "url": "https://canvas-mcp-sse.ariff.dev/sse"
    }
  }
}

Getting API Token

1. Log in to https://learn.mywhitecliffe.com
2. Go to Account → Settings
3. Scroll to "Approved Integrations"
4. Click "+ New Access Token"
5. Set expiry (recommended: end of semester)
6. Copy token immediately (shown only once)

Storing Token Securely

macOS Keychain:

security add-generic-password -a "$USER" -s "canvas-api" -w "YOUR_TOKEN"

Environment variable:

# Add to ~/.zshrc
export CANVAS_TOKEN="$(security find-generic-password -a "$USER" -s "canvas-api" -w)"

Important Notes

⚠️ Always verify with MCP tools - Don't assume course/assignment details

⚠️ Token security - Never commit tokens to git

⚠️ Rate limits - Canvas has API rate limits, batch requests when possible

⚠️ Course IDs change - Each semester has new course IDs

Troubleshooting

"Unauthorized" errors:

• Token expired - generate new one
• Token revoked - check Canvas settings

MCP not responding:

• Check https://canvas-mcp-sse.ariff.dev/health
• Verify mcp-config.json syntax

Missing courses:

• Check enrollment status
• Some courses may be unpublished