Autonomous Self-Healing CI/CD Pipeline
- Frontend: Vercel
- Backend: Render
graph LR
%% Styles
classDef user fill:#FFB6C1,stroke:#333,stroke-width:2px,color:black,font-weight:bold
classDef frontend fill:#E0F7FA,stroke:#00ACC1,stroke-width:2px,color:black,font-weight:bold
classDef backend fill:#E8F5E9,stroke:#43A047,stroke-width:2px,color:black,font-weight:bold
classDef ai fill:#F3E5F5,stroke:#8E24AA,stroke-width:2px,color:black,font-weight:bold
classDef ext fill:#FFF3E0,stroke:#FB8C00,stroke-width:2px,color:black
classDef db fill:#ECEFF1,stroke:#546E7A,stroke-width:2px,color:black
subgraph Canvas [Autonomous CI/CD Ecosystem]
direction LR
style Canvas fill:#f9f9f9,stroke:#333,stroke-width:1px,color:black
%% Nodes
User([👤 User]):::user
GitHub[🐱 GitHub Repo]:::ext
Gemini[🧠 Gemini AI]:::ai
subgraph Platform [Platform Internal]
direction TB
style Platform fill:#fff,stroke:#999,stroke-width:1px
Frontend[🖥️ Dashboard]:::frontend
Backend[⚙️ Backend API]:::backend
Results[(📊 Results)]:::db
subgraph Engine [Execution Engine]
style Engine fill:#F5F5F5,stroke:#ccc,stroke-width:1px
Orchestrator[🎼 Agents]:::backend
Sandbox[📦 Sandbox]:::ext
end
end
%% Connections within Canvas
User -->|1. URL| Frontend
Frontend -->|2. Start| Backend
Backend -->|3. Init| Orchestrator
Orchestrator <-->|4. Code| GitHub
Orchestrator <-->|5. Fix| Gemini
Orchestrator <-->|6. Test| Sandbox
Orchestrator -->|7. Data| Results
Results -->|8. View| Frontend
end
- Framework: React 19
- Build Tool: Vite 7
- Styling: TailwindCSS 4
- State Management: Zustand
- Animations: Framer Motion
- Visualization: Recharts
- Language: Python 3.12
- API Framework: FastAPI
- AI Framework: CrewAI
- LLM: Google Gemini 2.0 Flash
- Containerization: Docker (Sandbox Execution)
The agent autonomously detects and fixes the following Python error types:
- SYNTAX: Missing colons, unmatched parentheses, invalid syntax.
- INDENTATION: Inconsistent code alignment, mixed tabs/spaces.
- IMPORT: Missing modules, typo in package names, circular imports.
- TYPE_ERROR: Invalid data type operations (e.g.,
str+int). - LINTING: Unused imports, variable redefinitions (via AST analysis).
- LOGIC: Incorrect return values, off-by-one errors (detected via Pytest output).
- Python 3.12+
- Node.js 18+
- Docker Desktop (for sandboxed execution)
- Google Gemini API Key
- Navigate to the backend directory:
cd backend - Install dependencies:
pip install -r requirements.txt
- Create a
.envfile:cp .env.example .env # Edit .env and add your GEMINI_API_KEY - Start the server:
python main.py
- Navigate to the frontend directory:
cd frontend - Install dependencies:
npm install
- Start the development server:
npm run dev
Create a .env file in the backend/ directory with the following keys:
GEMINI_API_KEY=your_api_key_here
# Optional: ANTHROPIC_API_KEY=your_anthropic_key- Open the frontend at
http://localhost:5173. - Enter the GitHub Repository URL you want to fix (e.g., a repo with known bugs).
- Enter your Team Name and Leader Name.
- Click "Analyze Repository".
- The agent will:
- Clone the repository.
- Run tests to identify failures.
- Iteratively fix bugs (Syntax -> Import -> Logic, etc.).
- Push fixed code to a new branch:
TEAM_NAME_LEADER_NAME_AI_Fix.
- Complex Logic Bugs: Deep algorithmic flaws may require human intervention if the LLM cannot infer the correct logic from test cases alone.
- Execution Time: While fast (~20-30s), complex repos with many dependencies may take longer to install/test in the sandbox.
- Rate Limits: Heavy usage depends on the Gemini API quota.
| Role | Name |
|---|---|
| Team Leader | Ashraf Ahmed |
| Member | Purvesh Didpaye |
| Member | Prajwal Bahekar |
| Member | Riddhi Balapure |
Built for RIFT 2026 Hackathon