Building ethical, human-centered superintelligence to empower individuals and transform global systems.
This repository is the foundation for an agent-driven ecosystem that powers platforms like Opportunity Engine—a personal transformation engine that generates scholarships, career roadmaps, and custom resources based on user goals.
Our mission is to build modular, ethical superintelligence systems that:
- Align with human values
- Elevate underserved voices
- Automate opportunity matching
- Deliver explainable, scalable, and just outcomes
| Agent / Module | Purpose |
|---|---|
roadmap-agent |
Generates personalized milestone plans for users based on goals |
resume-agent |
Creates tailored resume/LinkedIn profiles |
opportunity-agent |
Matches grants, internships, schools, and jobs |
insight-agent |
Analyzes local/global data for contextual recommendations |
alignment-core |
Implements Constitutional AI principles and safety checks |
A UI-driven system that allows users to input:
- Name
- Zip Code
- Current Standing
- Dream Outcome
And receive:
- 📍 A step-by-step roadmap
- 📄 A customized resume
- 🔍 Curated opportunities
- 🤖 Ongoing AI support via chat agents
- Frontend: HTML/CSS, custom neural UI, React (future)
- Backend: Firebase (Firestore, Functions, Auth)
- AI: In-house models built on open-source frameworks (planned)
- Infra: Node.js, GitHub Actions, Firebase Hosting
This system follows ethical design guidelines including:
- Value alignment (human-in-the-loop feedback)
- Transparency (explainable agent logic)
- Global accessibility (multi-lingual + inclusive)
- Cultural adaptability
- Long-term safety and governance
Inspired by principles from:
- UNESCO AI Ethics
- Anthropic's Constitutional AI
- Future of Life Institute
⚠️ Setup instructions and sample data coming soon.
- Run the environment setup script:
./scripts/setupEnv.sh
- Install root dependencies:
npm install
- Install Cloud Functions dependencies:
npm install --prefix functions
- Install frontend dependencies:
Run this separately after the root install. The Cloud Build pipeline also performs
npm install --prefix frontend
npm installso the Vite production build works. - Create environment files in
frontend/using the provided example:Edit these files to include your Firebase config such ascp frontend/.env.example frontend/.env cp frontend/.env.example frontend/.env.local cp frontend/.env.production frontend/.env.production.local
VITE_FIREBASE_API_KEY. - Run tests from the repo root (requires functions dependencies):
npm test --silent - If the Firebase emulator reports authentication errors, re-authenticate using:
firebase login --reauth
To run Firebase Hosting locally and start the frontend app:
./start-dev.shIf you're not logged into Firebase, it will prompt you.
On CI, Cloud Build loads _FIREBASE_TOKEN from Secret Manager.
If the secret is not configured, Cloud Build skips deploying Firebase functions
and continues with the Cloud Run deploy.
GitHub Actions uses the same secret, or reads FIREBASE_TOKEN from a local
.env file when the secret isn't set and prints a warning otherwise.
- Generate a token locally using
firebase login:ci. - Store it as
firebase-ci-tokenin Secret Manager and grant Cloud Build access:gcloud secrets create firebase-ci-token --replication-policy=automatic printf '%s' "$FIREBASE_TOKEN" | \ gcloud secrets versions add firebase-ci-token --data-file=- gcloud secrets add-iam-policy-binding firebase-ci-token \ --member="serviceAccount:$PROJECT_ID@cloudbuild.gserviceaccount.com" \ --role="roles/secretmanager.secretAccessor"
- Pushes to
maintrigger Cloud Build, which installs all dependencies, runs tests, builds the frontend with Vite, and deploys Firebase functions when the secret token is present before updating the Cloud Run service withgcloud run deploy.
This project uses a custom cloudbuild.yaml to deploy both Firebase Functions
and Cloud Run services.
- Add the following secret to Secret Manager: Name: firebase-ci-token Value: [your Firebase CI token]
- Grant access to the Cloud Build service account.
- Link your GitHub repo to Google Cloud Build.
- Go to Cloud Build → Triggers.
- Click Create Trigger.
- Choose:
- Event: Push to a branch
- Branch:
main - Configuration: Cloud Build configuration file
- File location:
/cloudbuild.yaml - Service Account: Use custom role with only required permissions
- Save the trigger.
✅ Your deployments will now be fully managed through this YAML file.
Before deploying, create a production build of the React app. From the repository root run:
npm run buildThe command outputs static files in frontend/build/. Copy them to the repository's
public/ directory or set hosting.public in firebase.json to frontend/build so
Firebase Hosting serves the build directory. Cloud Build runs npm --prefix frontend run build
automatically during CI, but you can run it manually before:
firebase deployTo push the frontend and Cloud Functions together:
npm run build && firebase deployWhen deploying the Node server (for example via App Engine) you must provide
Firebase credentials for firebase-admin. Place a serviceAccount.json file in
the repository root or set the GOOGLE_APPLICATION_CREDENTIALS environment
variable to point at your service account file. The file name is already listed
in .gitignore so your credentials won't be committed.
Developers can review recent agent activity through /debug.html once
deployed. The page authenticates with Firebase Auth and checks an email
allowlist defined by functions.config().debug.allowlist. It fetches
log entries from the getLogs Cloud Function and presents them in a
filterable, auto-refreshing table grouped by agent.
/agent-hub.html offers a registry-driven view of all agents with links
to their docs and actions to re-run them. The registry is stored in
config/agents.json and mirrored under public/config/agents.json for
hosting.
A new runWebsiteAnalysis Cloud Function executes the website-analysis
flow ported from the ai-agent-systems repository. Trigger it with a
POST request containing the target URL:
curl -X POST https://<REGION>-<PROJECT>.cloudfunctions.net/runWebsiteAnalysis \
-H "Content-Type: application/json" \
-d '{"url":"https://example.com"}'The response contains the captured flow state and outputs for each step.
If Cloud Functions report Error: Cannot find module 'firebase-admin', install
dependencies explicitly:
cd functions
npm install firebase-admin
npm --prefix functions installEnsure firebase-admin appears in functions/package.json under
dependencies. Validate the installation with:
npm --prefix functions run build
timeout 3 npm --prefix functions start || trueIf the start command complains about missing firebase-admin, run these
commands again to ensure the dependency installed correctly.
Both commands should exit without errors.
When deployments fail, first confirm PROJECT_ID and other required
variables are defined in your Cloud Build trigger or GitHub workflow.
_FIREBASE_TOKEN is optional — if it's missing, the build skips the functions deploy step. Ensure the service account used has permission to deploy Firebase functions and Cloud Run.
Firebase functions and Cloud Run.
Run a local test build with:
gcloud builds submit --config cloudbuild.yamlLogs stream to Cloud Logging by default. If a logsBucket is configured,
logs will also appear in that Cloud Storage bucket.
- Project Name: Super Intelligence
- Project ID:
super-intelligence-7b653 - Project Number:
170923536461
This project is currently in active prototyping. If you're interested in building ethical agent systems, AI for public good, or opportunity pipelines for underserved communities, get in touch or open an issue.
We believe superintelligence should serve everyone—not just those with access or capital. This repo is our commitment to building systems that lift people into alignment with their highest potential.
"Ethical superintelligence isn't about replacing human ambition—it's about scaling it with justice."
© 2025 Csp-Ai • Licensed under MIT