optical-quantum-kernel

# Optical Quantum Kernel Skill

This skill simulates a photonic quantum computer that uses optical fibers for storage and linear optics for computation.
It calculates the quantum kernel (similarity) between two data vectors by encoding them into optical phases, passing them through simulated fibers (with loss), and interfering them.

## Security Features
- **Resource Bounding**: Capped at 8 modes to prevent resource exhaustion.
- **Input Validation**: Strict checks on input vector dimensions and limits.
- **Physics-Based Constraints**: Includes attenuation and phase noise for realism.

## Commands

- `simulate`: Run the quantum kernel simulation on two input vectors.

# Optical Quantum Kernel Simulation

This skill simulates a "Quantum Kernel Trick" using a physics-based model of an optical quantum computer.

## The Physics Model
1.  **Optical Fiber Storage**: The quantum state (photons) travels through optical fibers.
    -   **Loop**: Storage is simulated as a fiber loop.
    -   **Loss**: Attenuation ($0.2 dB/km$) reduces the signal amplitude.
    -   **Noise**: Random phase drift simulates thermal fluctuations in the fiber.
2.  **Linear Optics**: 
    -   **Encoding**: Data is mapped to the phase of optical modes.
    -   **Computation**: A 50:50 Beam Splitter interferes the modes.
3.  **Measurement**: Photodetectors measure the intensity at the output ports to estimate the kernel (similarity).

## Security Features
- **Bounded Resources**: The simulation acts as a "Sandbox", strictly creating only up to 8 optical modes/qubits. This prevents Denial of Service (DoS) via memory exhaustion.
- **Input Sanitization**: Verifies that inputs are valid numerical vectors before processing.

## Usage

Calculate the kernel (similarity) between two vectors:

```bash
python3 scripts/optical_kernel.py "0.5,1.2" "0.5,1.2"
# Expected: High kernel value (~1.0)
```

Orthogonal/Different vectors:
```bash
python3 scripts/optical_kernel.py "0.1,0.2" "0.9,0.8"
# Expected: Lower kernel value
```

## Publishing to ClawHub

To publish this skill:
```bash
clawhub publish
```