-
Notifications
You must be signed in to change notification settings - Fork 0
Consumer Setup
How to set up a downstream project to use Bluetemberg.
Add or create .npmrc in your project root:
@prototypdigital:registry=https://npm.pkg.github.com
//npm.pkg.github.com/:_authToken=${GITHUB_TOKEN}
Set GITHUB_TOKEN to a personal access token with read:packages scope.
For CI (GitHub Actions):
- uses: actions/setup-node@v4
with:
node-version: '20'
registry-url: 'https://npm.pkg.github.com'
scope: '@prototypdigital'
env:
NODE_AUTH_TOKEN: ${{ secrets.GITHUB_TOKEN }}npx @prototypdigital/bluetemberg initFollow the interactive prompts to select platforms, rules, agents, and skills.
Add to your GitHub Actions workflow to catch drift:
- name: Check AI config sync
run: npx bluetemberg sync --checkThis step exits with code 1 if any platform-specific file (.cursor/rules/, .claude/rules/, .github/instructions/, .cursor/mcp.json, .github/prompts/, etc.) is out of sync with the source in llm/ (and with optional adapters output, if you use them). It prevents the common pattern where someone edits a rule in llm/ and forgets to run sync before pushing. Without this check, the AI tools on different platforms silently diverge.
After editing any file in llm/:
# Regenerate platform files
npm run sync:llm-config
# Verify in CI
npm run sync:llm-config:checkTo get the latest starter templates and sync engine:
npm update @prototypdigital/bluetembergThen re-run sync to pick up any engine changes:
npx bluetemberg syncIf your project already has AI config files (.cursor/rules/, .claude/rules/, etc.) that were created manually:
- Run
npx @prototypdigital/bluetemberg initin your project - Move your existing rule content into
llm/rules/with the correct frontmatter format - Run
npx bluetemberg syncto regenerate platform files - Verify the output matches your previous setup
- Delete any manually-created platform files that are now generated