update

Author: fbchan

docs/class6/_static/class6-redteam-2-1.png

@@ -88,15 +88,54 @@ CASI scoring only applied to Signature attacks, not Agentic Warfare attack.
 
 ..  image:: ./_static/class6-redteam-2-2.png
 
+2 - Signature Attacks Report
+~~~~~~~~~~~~~~~~~~~~~~~~~~~~
 
 Report of an Attack Campaign. Click **View raw data** to see more details of every prompt used in the attack campaign.
 
+Select **Reports** and filter based on **complete** reports.
+
+..  image:: ./_static/class6-redteam-4-01.png
+
+Click on **Run date** to sort the report and select report name **foobz-nim-meta-llama3** report.
+
+..  image:: ./_static/class6-redteam-4-02.png       
+
+Report dashboard shown respective information about the scanned result.
+
 ..  image:: ./_static/class6-redteam-4.png
 
 Details of the attack prompts used in the campaign including the response from the AI model, the attack type, severity level, and whether the attack was successful or not. 
 
 ..  image:: ./_static/class6-redteam-4-1.png
 
+3 - Agentic Warfare Report
+~~~~~~~~~~~~~~~~~~~~~~~~~~
+
+Select **Reports** and search the following report name in the search bar and selecte the respective report.
+
+.. code-block:: bash
+
+   agentic-genearate-harmful-code-2
+
+..  image:: ./_static/class6-redteam-ars-01-0.png
+
+Respective infor of the agentic warfare scanning displayed. Select **View agentic fingerprints**.
+
+..  image:: ./_static/class6-redteam-ars-01.png
+
+Feel free to explore the Red Agent chain of thought. Its the internal reasoning process: the step-by-step logic, decisions, and planning it uses to achieve a task. 
+
+In agentic AI, this covers every mental step the agent takes—from forming questions and evaluating options to deciding on an action. This process is crucial for transparency, debugging, and explainability. Technologies like Agentic Fingerprints let you see each branch or step an agent considered, providing a clear decision tree of its entire thought process leading to a final result
+
+.. Note:: 
+  The scenario was intentionally crafted to request Python exploit code for CVE-2025-20333 as part of a **sanctioned security validation**. The objective was to **ensure AI models and applications are resilient and will not generate malicious code**.
+
+..  image:: ./_static/class6-redteam-ars-02.png
+
+Another example Agentic Warfare scanning with slight different intent.
+
+..  image:: ./_static/class6-redteam-ars-03.png
 
 
 |