Contributing to CoursDev AI

First-time setup

1. Clone the repo: git clone <repo-url> coursedev_ai
2. Open the extension/ folder in VS Code
3. Run npm install inside extension/
4. Press F5 to launch an Extension Development Host

Adding a skill

1. Copy extension/skills/custtr-example/ and rename it custtr-<your-skill-name>/
2. Edit SKILL.md — the name frontmatter field must match the folder name
3. Set version.txt to 1.0 (or increment if updating an existing skill)
4. Commit and open a PR — no other approval needed

Skills are named custtr-<name>. Examples: custtr-coursedev, custtr-pdf-helper, custtr-id.

Adding an MCP tool

1. Add your function to extension/docker/mcp-server/src/tools/custom.ts — or create a new file in that folder if the tool is large enough to warrant it
2. Register it in extension/docker/mcp-server/src/server.ts
3. Rebuild: npm run build inside extension/docker/mcp-server/
4. Test locally (see below), then open a PR

Testing locally

# Start the Docker container
cd extension/docker
docker compose up -d

# Verify the MCP server responds
docker compose exec coursedev-ai-tools node dist/server.js

Submitting a PR

• Branch from main
• PR description must state: which skill or tool changed, and how you tested it
• All team members can review and merge — there are no required reviewers

Releases

Maintainers push a v* tag (e.g., v1.1.0) to trigger CI. CI will automatically:

• Build and push the Docker image to AMD Artifactory
• Package and upload the VSIX to the GitHub Release
• Update releases/latest.json so installed extensions prompt users to update
• Bump the patch version in package.json for the next cycle