← Videos

01 AI Reconnaissance
AI Reconnaissance: What Is This Chatbot Connected To? | AI Security Testing 01
In this episode
0:00What is this chatbot connected to?0:32Application and gateway1:01Orchestration, retrieval and tools1:44Inference runtime and model2:18Passive and active reconnaissance3:08HEAD requests and header clues3:55From page to script to API route4:39Send and preserve a normal request5:31Compare general and document questions6:22Map evidence, permissions and unknowns7:02Model fingerprinting: limits of clues7:40Choose the next test and optional practice
Chapter times come from the published description.
From the episode description
AI Reconnaissance: What Is This Chatbot Connected To? | AI Security Testing 01
Learn to map an AI application before choosing a security test. We start with its user interface, gateway, orchestration, retrieval, tools, integrations, inference runtime and model.
Then follow concrete curl examples: inspect response headers, follow a page's JavaScript to an assistant route, send an ordinary JSON request, and save its headers and answer. Compare general and document-specific questions, distinguish clues from evidence, and turn the resulting map into a focused next test.
HarborDesk, its address, requests and responses are original mock examples. No live target or P2501 result is claimed. The commands demonstrate practical syntax; use the actual address, request schema and normal authorized access of your practice system. Reported sources alone do not prove how retrieval ran.
Optional practice: P2501 Lab 2.1, Laughing Man. P2501 is not required to follow the lesson.
https://p2501.ai
New lessons each week, progressing from reconnaissance to prompt injection, tools, retrieval, memory, supply chains and retesting.
curl syntax reference: https://curl.se/docs/manpage.html
#AISecurity #Reconnaissance #ResearchNode
Chapters
0:00 What is this chatbot connected to?
0:32 Application and gateway
1:01 Orchestration, retrieval and tools
1:44 Inference runtime and model
2:18 Passive and active reconnaissance
3:08 HEAD requests and header clues
3:55 From page to script to API route
4:39 Send and preserve a normal request
5:31 Compare general and document questions
6:22 Map evidence, permissions and unknowns
7:02 Model fingerprinting: limits of clues
7:40 Choose the next test and optional practice
Descriptions and chapter markers are supplied by the channel. They are not a transcript or an independent check of every claim in the video.