AI Advisor

Interactive AI-powered testing assistant with conversation history.

Table of contents

1. Overview
2. How to Use
   1. Sending a Request
   2. Asking Follow-Up Questions
   3. Attaching Additional Requests
3. Conversation Flow
4. Using Templates with AI Advisor
5. Session Management
6. Context Variables
7. Tips for Best Results

---

Overview

The AI Advisor is your interactive security testing companion. Send any HTTP request to the AI Advisor, and it provides targeted vulnerability analysis, testing payloads, and step-by-step exploitation guidance. It maintains full conversation history so you can ask follow-up questions.

---

How to Use

Sending a Request

From anywhere in Burp Suite:

1. Right-click any HTTP request (Proxy, Target, Repeater, etc.)
2. Select “💡 Send to VISTA AI Advisor”
3. VISTA switches to the AI Advisor tab
4. AI automatically analyzes the request and provides:
   • Vulnerability assessment
   • Testing payloads
   • Step-by-step methodology
   • Risk rating

Asking Follow-Up Questions

After the initial analysis, type follow-up questions in the chat input:

• “Can you provide more XSS payloads for this parameter?”
• “How would I bypass the WAF for this endpoint?”
• “What about blind SQL injection?”
• “Show me the Turbo Intruder script for this race condition”

The AI remembers the full conversation context and the attached request.

Attaching Additional Requests

For multi-request analysis (e.g., testing a workflow):

1. Right-click another request in Burp
2. Select “📎 Attach to Interactive Assistant”
3. The request is added to the current conversation context
4. AI can now analyze the relationship between multiple requests

---

Conversation Flow

┌──────────────────────────────────────────────┐
│  1. User sends request via right-click       │
│     → "💡 Send to VISTA AI Advisor"          │
└──────────────────┬───────────────────────────┘
                   │
┌──────────────────▼───────────────────────────┐
│  2. VISTA builds analysis context:           │
│     • HTTP request/response data             │
│     • Parameter analysis                     │
│     • WAF detection results                  │
│     • Selected template (if any)             │
│     • Variable context (35 variables)        │
└──────────────────┬───────────────────────────┘
                   │
┌──────────────────▼───────────────────────────┐
│  3. AI Provider processes prompt             │
│     (OpenAI / Azure / OpenRouter)            │
└──────────────────┬───────────────────────────┘
                   │
┌──────────────────▼───────────────────────────┐
│  4. Response displayed in chat UI            │
│     • Markdown rendering                     │
│     • Code blocks with copy button           │
│     • Structured analysis                    │
└──────────────────┬───────────────────────────┘
                   │
┌──────────────────▼───────────────────────────┐
│  5. User asks follow-up (conversation mode)  │
│     • Full history maintained                │
│     • Context-aware responses                │
│     • Attach more requests if needed         │
└──────────────────────────────────────────────┘

---

Using Templates with AI Advisor

Templates dramatically improve AI response quality by providing structured, expert-level prompts:

1. Go to Prompt Templates tab
2. Select a template (e.g., “SSRF (Expert)”)
3. Click “Use Template”
4. Now send a request to AI Advisor — the template shapes the analysis
5. AI provides deep, methodology-driven guidance specific to that vulnerability class

Always use an expert template when you know what vulnerability class you’re testing. The difference in response quality is significant.

---

Session Management

• Conversations persist across interactions within the same session
• New request starts a fresh session (previous session is properly closed)
• Sessions are saved to ~/.vista/sessions/ for persistence across Burp restarts
• Thread-safe — conversation history is synchronized for safe concurrent access

---

Context Variables

The AI Advisor automatically extracts and injects 35 context variables from the HTTP request:

Variable	Description
``	Full raw HTTP request
``	Full raw HTTP response
``	Request URL
``	HTTP method (GET, POST, etc.)
``	All parameters with values
``	Request headers
``	Request body
``	Detected endpoint type (API, form, etc.)
``	Detected WAF information
``	Calculated risk score
``	Deep analysis of request patterns
``	Deep analysis of response patterns
``	Detected sensitive data in traffic
``	Error messages found in response
``	User’s question text

These variables are automatically populated and used by templates to provide rich, context-aware analysis.

---

Tips for Best Results

Be specific — Instead of “find vulnerabilities,” ask “test the search parameter for reflected XSS with WAF bypass.”

Use follow-ups — After initial analysis, drill down: “show me payloads for the second vulnerability you identified.”

Attach related requests — When testing multi-step flows (login → token → API call), attach all related requests for holistic analysis.