# Microsoft AutoGen Integration
[](https://github.com/sweeden-ttu/canvas-lms-mcp/actions/workflows/autogen-ci.yml)
[](https://gitlab.com/sweeden-ttu/canvas-lms-mcp)
[](https://www.python.org/)
[](https://microsoft.github.io/autogen/)
This guide covers integrating the Canvas LMS MCP Server with Microsoft AutoGen for multi-agent AI workflows, including CI/CD pipelines and GitHub/GitLab synchronization.
---
## Table of Contents
- [Overview](#overview)
- [Prerequisites](#prerequisites)
- [Installation](#installation)
- [AutoGen Agent Configuration](#autogen-agent-configuration)
- [CI/CD Integration](#cicd-integration)
- [GitHub/GitLab Synchronization](#githubgitlab-synchronization)
- [Badge Generation](#badge-generation)
- [Troubleshooting](#troubleshooting)
---
## Overview
Microsoft AutoGen enables building multi-agent systems where AI agents collaborate to solve complex tasks. This integration allows:
- **Canvas MCP Agent**: Interfaces with Canvas LMS via the MCP server
- **CI/CD Agent**: Manages build, test, and deployment pipelines
- **Sync Agent**: Keeps GitHub and GitLab repositories synchronized
- **Badge Agent**: Generates status badges and thumbnails
### Architecture
```
┌─────────────────────────────────────────────────────────────┐
│ AutoGen Orchestrator │
├─────────────┬─────────────┬─────────────┬───────────────────┤
│ Canvas MCP │ CI/CD │ Sync │ Badge │
│ Agent │ Agent │ Agent │ Agent │
├─────────────┼─────────────┼─────────────┼───────────────────┤
│ Canvas LMS │ GitHub │ GitHub ↔ │ shields.io │
│ API │ Actions │ GitLab │ badgen.net │
└─────────────┴─────────────┴─────────────┴───────────────────┘
```
---
## Prerequisites
- Python 3.10+
- uv package manager
- GitHub account with Actions enabled
- GitLab account (for sync)
- Canvas API token
- OpenAI or Azure OpenAI API key (for AutoGen LLM)
---
## Installation
### 1. Install AutoGen Dependencies
AutoGen 0.3+ uses the `autogen-agentchat` and `autogen-ext` packages (not `pyautogen`). Install as follows:
```bash
cd canvas-lms-mcp
# AutoGen 0.3+ AgentChat and OpenAI model client
uv add "autogen-agentchat" "autogen-ext[openai]" "openai>=1.0.0"
# Optional: Azure OpenAI with AAD
uv add "autogen-ext[azure]" "azure-identity>=1.15.0"
```
Note: If you are on AutoGen 0.2, use `pyautogen` and the legacy `from autogen import ...` API; see the [AutoGen migration guide](https://microsoft.github.io/autogen/stable/user-guide/agentchat-user-guide/migration-guide.html) for 0.2 to 0.3.
### 2. Configure Environment
Add to your `.env` file:
```env
# Canvas LMS Configuration
CANVAS_API_TOKEN=your_canvas_api_token_here
CANVAS_BASE_URL=https://texastech.instructure.com
# AutoGen LLM Configuration
OPENAI_API_KEY=your_openai_api_key_here
# OR for Azure OpenAI:
AZURE_OPENAI_API_KEY=your_azure_key_here
AZURE_OPENAI_ENDPOINT=https://your-resource.openai.azure.com/
AZURE_OPENAI_DEPLOYMENT=gpt-4
# GitLab Sync Configuration
GITLAB_TOKEN=your_gitlab_token_here
GITLAB_PROJECT_ID=your_project_id
```
---
## AutoGen Agent Configuration
### Create Agent Configuration File
**`autogen_config.py`** (AutoGen 0.3+ API using `autogen_agentchat` and `autogen_ext`):
```python
"""AutoGen agent configuration for Canvas LMS MCP integration (AutoGen 0.3+)."""
import asyncio
import os
from autogen_agentchat.agents import AssistantAgent, UserProxyAgent
from autogen_agentchat.conditions import TextMentionTermination
from autogen_agentchat.teams import RoundRobinGroupChat
from autogen_ext.models.openai import OpenAIChatCompletionClient
# Model client for OpenAI (required by AssistantAgent in 0.3+)
model_client = OpenAIChatCompletionClient(
model="gpt-4",
api_key=os.getenv("OPENAI_API_KEY"),
)
# Canvas MCP Agent - Interfaces with Canvas LMS
canvas_agent = AssistantAgent(
name="CanvasMCPAgent",
model_client=model_client,
system_message="""You are an expert at interfacing with Canvas LMS via the MCP server.
You can:
- List courses and enrollments
- Retrieve assignments, modules, and grades
- Fetch announcements and discussion topics
- Access calendar events and planner items
Use the MCP tools available to query Canvas data. Always format responses
clearly and handle errors gracefully.""",
)
# CI/CD Agent - Manages build and deployment pipelines
cicd_agent = AssistantAgent(
name="CICDAgent",
model_client=model_client,
system_message="""You are an expert CI/CD engineer specializing in GitHub Actions and GitLab CI.
You can:
- Create and modify workflow YAML files
- Configure build, test, and deployment stages
- Set up environment secrets and variables
- Debug pipeline failures
- Generate status badges
Follow best practices for CI/CD security and efficiency.""",
)
# Sync Agent - Synchronizes GitHub and GitLab repositories
sync_agent = AssistantAgent(
name="SyncAgent",
model_client=model_client,
system_message="""You are an expert at repository synchronization between GitHub and GitLab.
You can:
- Configure bidirectional mirroring
- Handle merge conflicts during sync
- Set up webhooks for real-time synchronization
- Manage branch protection rules on both platforms
- Handle large file storage (LFS) synchronization
Ensure data integrity and maintain commit history during sync operations.""",
)
# Badge Agent - Generates status badges and thumbnails
badge_agent = AssistantAgent(
name="BadgeAgent",
model_client=model_client,
system_message="""You are an expert at generating status badges and thumbnails for repositories.
You can:
- Create shields.io badge URLs
- Generate custom badge images
- Configure dynamic badges from CI/CD status
- Create workflow status badges
- Design repository thumbnail images
Always provide markdown-ready badge code. Reply with TERMINATE when done.""",
)
# User Proxy Agent - Represents the user in the conversation
user_proxy = UserProxyAgent(name="UserProxy")
# Termination when any agent replies with "TERMINATE"
termination = TextMentionTermination("TERMINATE")
# Team (replaces GroupChat + GroupChatManager in 0.2)
team = RoundRobinGroupChat(
participants=[user_proxy, canvas_agent, cicd_agent, sync_agent, badge_agent],
termination_condition=termination,
max_turns=20,
)
async def run_task_async(task: str):
"""Run a task using the agent team. Returns TaskResult with messages."""
result = await team.run(task=task)
return result.messages
def run_task(task: str):
"""Run a task using the agent team (sync wrapper)."""
return asyncio.run(run_task_async(task))
```
### Usage Example
```python
from autogen_config import run_task
# Example: Sync repos and generate badges (run_task returns list of messages)
messages = run_task("""
1. Check the current GitHub Actions workflow status
2. Sync the repository to GitLab
3. Generate updated status badges for:
- Build status
- Test coverage
- GitLab sync status
4. Update the README with the new badges. Reply with TERMINATE when done.
""")
```
---
## CI/CD Integration
### GitHub Actions Workflow
Create **`.github/workflows/autogen-ci.yml`**:
```yaml
name: AutoGen CI/CD Pipeline
on:
push:
branches: [main, develop]
pull_request:
branches: [main]
workflow_dispatch:
env:
PYTHON_VERSION: "3.11"
UV_CACHE_DIR: /tmp/.uv-cache
jobs:
# ============================================
# Build and Test Stage
# ============================================
build-test:
name: Build & Test
runs-on: ubuntu-latest
steps:
- name: Checkout repository
uses: actions/checkout@v4
with:
fetch-depth: 0 # Full history for sync
- name: Set up Python
uses: actions/setup-python@v5
with:
python-version: ${{ env.PYTHON_VERSION }}
- name: Install uv
uses: astral-sh/setup-uv@v4
with:
enable-cache: true
cache-dependency-glob: "uv.lock"
- name: Install dependencies
run: uv sync --all-extras --dev
- name: Run linting
run: uv run ruff check .
- name: Run type checking
run: uv run mypy server.py
- name: Run tests
run: uv run pytest tests/ -v --tb=short
env:
CANVAS_API_TOKEN: ${{ secrets.CANVAS_API_TOKEN }}
CANVAS_BASE_URL: ${{ secrets.CANVAS_BASE_URL }}
- name: Upload coverage
uses: codecov/codecov-action@v4
with:
token: ${{ secrets.CODECOV_TOKEN }}
fail_ci_if_error: false
# ============================================
# GitLab Sync Stage
# ============================================
gitlab-sync:
name: Sync to GitLab
runs-on: ubuntu-latest
needs: build-test
if: github.ref == 'refs/heads/main'
steps:
- name: Checkout repository
uses: actions/checkout@v4
with:
fetch-depth: 0
- name: Push to GitLab
run: |
git remote add gitlab https://oauth2:${{ secrets.GITLAB_TOKEN }}@gitlab.com/${{ secrets.GITLAB_PROJECT_PATH }}.git || true
git push gitlab main --force
git push gitlab --tags --force
env:
GITLAB_TOKEN: ${{ secrets.GITLAB_TOKEN }}
- name: Notify sync complete
run: |
echo "::notice::Successfully synced to GitLab"
# ============================================
# AutoGen Agent Stage
# ============================================
autogen-agents:
name: Run AutoGen Agents
runs-on: ubuntu-latest
needs: build-test
if: github.event_name == 'workflow_dispatch'
steps:
- name: Checkout repository
uses: actions/checkout@v4
- name: Set up Python
uses: actions/setup-python@v5
with:
python-version: ${{ env.PYTHON_VERSION }}
- name: Install uv
uses: astral-sh/setup-uv@v4
- name: Install AutoGen dependencies
run: |
uv sync
uv add autogen-agentchat "autogen-ext[openai]" openai
- name: Run AutoGen pipeline
run: |
uv run python -c "
from autogen_config import run_task
run_task('Generate status report for CI/CD pipeline. Reply with TERMINATE when done.')
"
env:
OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}
CANVAS_API_TOKEN: ${{ secrets.CANVAS_API_TOKEN }}
# ============================================
# Badge Generation Stage
# ============================================
generate-badges:
name: Generate Status Badges
runs-on: ubuntu-latest
needs: [build-test, gitlab-sync]
if: always()
steps:
- name: Checkout repository
uses: actions/checkout@v4
- name: Generate badge JSON
run: |
mkdir -p .github/badges
# Build status badge
if [ "${{ needs.build-test.result }}" == "success" ]; then
echo '{"schemaVersion":1,"label":"build","message":"passing","color":"brightgreen"}' > .github/badges/build.json
else
echo '{"schemaVersion":1,"label":"build","message":"failing","color":"red"}' > .github/badges/build.json
fi
# Sync status badge
if [ "${{ needs.gitlab-sync.result }}" == "success" ]; then
echo '{"schemaVersion":1,"label":"GitLab Sync","message":"synced","color":"orange"}' > .github/badges/sync.json
else
echo '{"schemaVersion":1,"label":"GitLab Sync","message":"pending","color":"yellow"}' > .github/badges/sync.json
fi
- name: Commit badge updates
run: |
git config user.name "github-actions[bot]"
git config user.email "github-actions[bot]@users.noreply.github.com"
git add .github/badges/
git diff --staged --quiet || git commit -m "chore: update status badges [skip ci]"
git push
```
### GitLab CI Configuration
Create **`.gitlab-ci.yml`** for GitLab-side pipeline:
```yaml
stages:
- build
- test
- sync
- badges
variables:
PYTHON_VERSION: "3.11"
PIP_CACHE_DIR: "$CI_PROJECT_DIR/.cache/pip"
cache:
paths:
- .cache/pip
- .venv/
# ============================================
# Build Stage
# ============================================
build:
stage: build
image: python:${PYTHON_VERSION}
script:
- pip install uv
- uv sync --all-extras --dev
artifacts:
paths:
- .venv/
expire_in: 1 hour
# ============================================
# Test Stage
# ============================================
test:
stage: test
image: python:${PYTHON_VERSION}
dependencies:
- build
script:
- pip install uv
- uv run pytest tests/ -v
coverage: '/TOTAL.*\s+(\d+%)$/'
lint:
stage: test
image: python:${PYTHON_VERSION}
dependencies:
- build
script:
- pip install uv
- uv run ruff check .
- uv run mypy server.py
allow_failure: true
# ============================================
# GitHub Sync Stage
# ============================================
sync-to-github:
stage: sync
image: alpine/git
only:
- main
script:
- git remote add github https://oauth2:${GITHUB_TOKEN}@github.com/${GITHUB_REPO}.git || true
- git push github HEAD:main --force
- git push github --tags --force
when: manual
# ============================================
# Badge Generation
# ============================================
generate-badges:
stage: badges
image: alpine
script:
- mkdir -p public/badges
- |
cat > public/badges/gitlab-build.json << EOF
{"schemaVersion":1,"label":"GitLab Build","message":"passing","color":"fc6d26"}
EOF
artifacts:
paths:
- public/badges/
```
---
## GitHub/GitLab Synchronization
### Sync Agent Implementation
**`sync_agent.py`:**
```python
"""Repository synchronization agent for GitHub and GitLab."""
import os
import subprocess
from dataclasses import dataclass
from typing import Optional
import httpx
@dataclass
class SyncConfig:
"""Configuration for repository synchronization."""
github_repo: str
github_token: str
gitlab_project_id: str
gitlab_token: str
gitlab_url: str = "https://gitlab.com"
class RepositorySyncAgent:
"""Agent for synchronizing repositories between GitHub and GitLab."""
def __init__(self, config: SyncConfig):
self.config = config
self.github_api = "https://api.github.com"
async def sync_github_to_gitlab(self, branch: str = "main") -> dict:
"""Sync GitHub repository to GitLab."""
result = {"status": "pending", "details": []}
try:
# Clone from GitHub
clone_url = f"https://oauth2:{self.config.github_token}@github.com/{self.config.github_repo}.git"
subprocess.run(
["git", "clone", "--mirror", clone_url, "repo_mirror"],
check=True,
capture_output=True
)
result["details"].append("Cloned from GitHub")
# Push to GitLab
gitlab_url = f"https://oauth2:{self.config.gitlab_token}@gitlab.com/{self.config.gitlab_project_id}.git"
subprocess.run(
["git", "-C", "repo_mirror", "push", "--mirror", gitlab_url],
check=True,
capture_output=True
)
result["details"].append("Pushed to GitLab")
result["status"] = "success"
except subprocess.CalledProcessError as e:
result["status"] = "failed"
result["error"] = str(e)
finally:
# Cleanup
subprocess.run(["rm", "-rf", "repo_mirror"], capture_output=True)
return result
async def get_sync_status(self) -> dict:
"""Get current synchronization status."""
async with httpx.AsyncClient() as client:
github_sha = "unknown"
gitlab_sha = "unknown"
error = None
try:
# Check GitHub
github_resp = await client.get(
f"{self.github_api}/repos/{self.config.github_repo}/commits/main",
headers={"Authorization": f"Bearer {self.config.github_token}"},
)
github_resp.raise_for_status()
github_sha = github_resp.json().get("sha", "unknown")[:7]
except (httpx.HTTPStatusError, ValueError) as e:
error = f"github: {e!s}"
try:
# Check GitLab
gitlab_resp = await client.get(
f"{self.config.gitlab_url}/api/v4/projects/{self.config.gitlab_project_id}/repository/commits/main",
headers={"PRIVATE-TOKEN": self.config.gitlab_token},
)
gitlab_resp.raise_for_status()
gitlab_sha = gitlab_resp.json().get("id", "unknown")[:7]
except (httpx.HTTPStatusError, ValueError) as e:
error = f"{error or ''}; gitlab: {e!s}".lstrip("; ")
return {
"github_sha": github_sha,
"gitlab_sha": gitlab_sha,
"in_sync": github_sha == gitlab_sha and error is None,
"error": error,
}
def generate_sync_badge(self, in_sync: bool) -> str:
"""Generate markdown badge for sync status."""
if in_sync:
return ""
else:
return ""
```
### Webhook Configuration
**GitHub Webhook Setup:**
1. Go to Repository Settings > Webhooks > Add webhook
2. Payload URL: `https://gitlab.com/api/v4/projects/{PROJECT_ID}/mirror/pull`
3. Content type: `application/json`
4. Events: Push, Release
**GitLab Mirroring Setup:**
1. Go to Project Settings > Repository > Mirroring repositories
2. Git repository URL: `https://github.com/{OWNER}/{REPO}.git`
3. Mirror direction: Pull
4. Authentication method: Password (use GitHub PAT)
---
## Badge Generation
### Available Badge Types
| Badge Type | Markdown | Preview |
|------------|----------|---------|
| Build Status | `` |  |
| GitLab Sync | `` |  |
| AutoGen | `` |  |
| Python | `` |  |
| Coverage | `` |  |
### Badge Generator Script
**`scripts/generate_badges.py`:**
```python
"""Generate status badges for the repository."""
import json
from pathlib import Path
from dataclasses import dataclass
from typing import Literal
@dataclass
class Badge:
"""Represents a status badge."""
label: str
message: str
color: str
logo: str = ""
def to_shields_url(self) -> str:
"""Generate shields.io URL."""
base = f"https://img.shields.io/badge/{self.label}-{self.message}-{self.color}"
if self.logo:
base += f"?logo={self.logo}"
return base.replace(" ", "%20")
def to_markdown(self, alt_text: str = "") -> str:
"""Generate markdown image."""
alt = alt_text or self.label
return f"})"
def to_json(self) -> dict:
"""Generate endpoint JSON for dynamic badges."""
return {
"schemaVersion": 1,
"label": self.label,
"message": self.message,
"color": self.color,
}
def generate_all_badges(output_dir: Path) -> list[str]:
"""Generate all project badges."""
output_dir.mkdir(parents=True, exist_ok=True)
badges = [
Badge("build", "passing", "brightgreen", "github"),
Badge("AutoGen", "enabled", "purple", "microsoft"),
Badge("GitLab", "synced", "orange", "gitlab"),
Badge("Python", "3.10+", "blue", "python"),
Badge("MCP", "active", "green"),
]
markdown_lines = []
for badge in badges:
# Save JSON endpoint
json_path = output_dir / f"{badge.label.lower()}.json"
json_path.write_text(json.dumps(badge.to_json(), indent=2))
# Collect markdown
markdown_lines.append(badge.to_markdown())
return markdown_lines
if __name__ == "__main__":
badges_dir = Path(".github/badges")
markdown = generate_all_badges(badges_dir)
print("Badge Markdown:")
print(" ".join(markdown))
```
---
## Troubleshooting
### Common Issues
**AutoGen connection timeout:**
```python
# Increase timeout in LLM config
LLM_CONFIG = {
"timeout": 300, # 5 minutes
"max_retries": 3,
}
```
**GitLab sync authentication failure:**
```bash
# Verify GitLab token has write_repository scope
curl -H "PRIVATE-TOKEN: $GITLAB_TOKEN" \
"https://gitlab.com/api/v4/projects/$PROJECT_ID"
```
**GitHub Actions secrets not available:**
- Ensure secrets are set in repository Settings > Secrets and variables > Actions
- Check workflow has correct permissions in `GITHUB_TOKEN` scope
### Logs and Debugging
```bash
# View AutoGen agent logs
export AUTOGEN_LOG_LEVEL=DEBUG
uv run python autogen_config.py
# Test GitLab sync manually
git remote add gitlab https://oauth2:$GITLAB_TOKEN@gitlab.com/user/repo.git
git push gitlab main --dry-run
```
---
## Related Documentation
- [Microsoft AutoGen Documentation](https://microsoft.github.io/autogen/)
- [GitHub Actions Documentation](https://docs.github.com/en/actions)
- [GitLab CI/CD Documentation](https://docs.gitlab.com/ee/ci/)
- [Shields.io Badge Service](https://shields.io/)
- [Canvas LMS MCP Server](../README.md)
