AgentSec Advisor

Ulinzi wa Mawakala wa AI

Secure your agentic AI systems with expert threat modeling

AgentSec Advisor

Ulinzi wa Mawakala wa AI
0 tok · $0.0000
Select Model

Compare mode — both models receive the same prompt

Model A

Karibu! Welcome to AgentSec Advisor 🛡️

I help you threat model agentic AI systems using the AWS Agentic AI Security Scoping Matrix.

Select a model above, or hit Compare Models to run your prompt side by side.

Responses will appear here when you send a prompt.
0/10000
Session
Tokens 0
Cost $0.0000
Requests 0
Last Response
Compare
Model A Model B
TTFT——
Total——
In tok——
Out tok——
Cost——

Document Q&A

Upload a PDF and ask security questions about it for up to 5 minutes.

Drop PDF here or

Up to ~10 MB
Session: 5:00
Session expired — upload a new document to continue.
Prompt Cache
Status —
Read tokens —
Write tokens —
TTFT —
Total —
Cost —

Document Q&A 📄

Upload a PDF on the left, then ask questions about it. The document is sent once and cached — subsequent questions reuse the cached context, saving tokens and reducing latency.

Supports Claude Haiku 4.5 (Anthropic prompt caching) and Nova Lite (Amazon Nova system caching).

Model Evaluation

3 standardised security prompts, AI-judged on Accuracy, Relevance, and Completeness (1–5).

Models — click to toggle:
Running evaluation…