This is an interactive demonstration of the SENTINEL (System for Elicitation Neutralization, Threat Interception, and Narrative Evaluation Logic) architecture. Enter a potentially malicious prompt below to see how each security layer processes it.
The following sections detail the logic for each layer of the SENTINEL system, presented in Python for conceptual clarity. The interactive demo above uses a JavaScript implementation of these principles.
... (code for sanitizer.py) ...... (code for heuristic_analyzer.py) ...... (code for meta_wrapper.py) ...Layers like an LLM Adjudicator and an Output Filter are powerful but cannot be simulated in a simple front-end environment as they require backend LLM API calls.