Build a multi-layered prompt injection defense system with input classification, output filtering, and canary token detection for LLM-powered applications.
Replace parameters with your LLM provider and application type. Start with Layer 1 classification and Layer 2 prompt protection. Add output validation and monitoring iteratively.
Initial release
Sign in and download this prompt to leave a review.