Settings Reference

This page documents every configurable setting in the extension, organized by section. Settings are grouped into tabs in the bottom settings panel of the AI Agent UI.

AI Backend

Setting	Type	Default	Description
Preferred Backend	Dropdown	codex-cli	Backend used for new chat sessions. The list is filtered to backends that are available on this machine.
Codex CLI Command	Text	codex chat	Shell command to launch the Codex backend (Windows npm shim example: C:\\Users\\<you>\\AppData\\Roaming\\npm\\codex.cmd).
Gemini CLI Command	Text	gemini --output-format text --model gemini-2.5-flash --yolo	Shell command to launch the Gemini backend (Windows npm shim example: C:\\Users\\<you>\\AppData\\Roaming\\npm\\gemini.cmd).
Claude CLI Command	Text	claude	Shell command to launch the Claude backend (Windows npm shim example: C:\\Users\\<you>\\AppData\\Roaming\\npm\\claude.cmd).
OpenCode CLI Command	Text	opencode	Shell command to launch the OpenCode backend (Windows npm shim example: C:\\Users\\<you>\\AppData\\Roaming\\npm\\opencode.cmd).
Ollama URL	Text	http://127.0.0.1:11434	Base URL for the Ollama HTTP API.
Ollama Model	Text	llama3.1	Model name to use with Ollama.
Ollama API Key	Text	(empty)	Optional Authorization: Bearer token for Ollama-compatible servers.
Ollama Extra Headers	Multiline	(empty)	Extra headers, one per line (Header: value).
Ollama Auto-Start	Toggle	On	Automatically start ollama serve when backend is selected.
Ollama Serve Command	Text	ollama serve	Custom command to start the Ollama server.
Ollama Timeout	Number	120	Timeout in seconds for Ollama requests (range: 30–3600).
Ollama Context Window	Number	8192	Context window sent to Ollama (range: 2048–128000).
LM Studio URL	Text	http://127.0.0.1:1234	Base URL for the LM Studio HTTP API.
LM Studio Model	Text	lmstudio	Model identifier for LM Studio.
LM Studio API Key	Text	(empty)	Optional Authorization: Bearer token for OpenAI-compatible servers.
LM Studio Extra Headers	Multiline	(empty)	Extra headers, one per line (Header: value).
LM Studio Auto-Start	Toggle	On	Automatically start LM Studio server.
LM Studio Server Command	Text	lms server start	Custom command to launch LM Studio server.
LM Studio Timeout	Number	120	Timeout in seconds for LM Studio requests (range: 30–3600).
OpenAI-Compatible URL	Text	(empty)	Base URL for an OpenAI-compatible provider (include /v1 or /vN if required).
OpenAI-Compatible Model	Text	(empty)	Model identifier sent in requests.
OpenAI-Compatible API Key	Text	(empty)	Optional Authorization: Bearer token.
OpenAI-Compatible Extra Headers	Multiline	(empty)	Extra headers, one per line (Header: value).
OpenAI-Compatible Timeout	Number	120	Timeout in seconds for OpenAI-compatible requests (range: 30–3600).
Auto-Restart	Toggle	On	Automatically restart a crashed CLI backend.
Agent Profile	Dropdown	pentester	System instruction profile. Options are loaded from ~/.burp-ai-agent/AGENTS/*.md (use Refresh to reload). See Agent Profiles.

Privacy & Logging

Setting	Type	Default	Description
Privacy Mode	Dropdown	OFF	Controls data redaction. Options: STRICT, BALANCED, OFF. See Privacy Modes.
Determinism Mode	Toggle	Off	Stabilizes context ordering for reproducible prompt bundles.
Host Anonymization Salt	Text	(auto-generated)	Secret string used to generate stable host pseudonyms in STRICT mode. Rotate per engagement.
Audit Logging	Toggle	Off	Enables JSONL audit logging of all AI interactions. Logs saved to ~/.burp-ai-agent/audit.jsonl.

MCP Server

Setting	Type	Default	Description
Enable MCP	Toggle	On	Start the MCP SSE server on extension load.
Host	Text	127.0.0.1	Bind address. Use 127.0.0.1 for local-only access.
Port	Number	9876	TCP port for the SSE server.
External Access	Toggle	Off	Allow connections from non-loopback addresses. Requires TLS.
STDIO Bridge	Toggle	Off	Enable stdin/stdout MCP transport in addition to SSE.
Token	Text	(auto-generated)	Bearer token for external access. Localhost access does not require a token.
TLS Enabled	Toggle	Off	Enable HTTPS for the MCP server.
Auto-Generate Certificate	Toggle	On	Generate a self-signed PKCS12 keystore automatically. Stored at ~/.burp-ai-agent/certs/mcp-keystore.p12.
Keystore Path	Text	(auto-populated)	Path to a custom PKCS12 keystore for TLS.
Keystore Password	Text	(auto-generated)	Password for the custom keystore.
Max Concurrent Requests	Number	4	Maximum parallel MCP tool calls (range: 1–64).
Max Body Bytes	Number	2097152	Maximum response body size returned by MCP tools (range: 256 KB – 100 MB).
Enable Unsafe Tools	Toggle	Off	Master switch for all tools marked as "unsafe".

Burp Integration

Setting	Type	Default	Description
Tool Toggles	Checkboxes	(varies)	Enable/disable individual MCP tools by category. See Tools Reference.

Passive AI Scanner

Setting	Type	Default	Description
Enabled	Toggle	Off	Enable background passive analysis of proxy traffic.
Rate Limit	Number	5	Minimum seconds between analysis requests (range: 1–60).
Scope Only	Toggle	On	Only analyze requests to in-scope targets.
Max Size (KB)	Number	96	Maximum request+response size sent for analysis (range: 16–1024 KB).
Min Severity	Dropdown	LOW	Ignore findings below this severity level. Options: LOW, MEDIUM, HIGH, CRITICAL.

Active AI Scanner

Setting	Type	Default	Description
Enabled	Toggle	Off	Enable active vulnerability testing.
Max Concurrent Scans	Number	3	Maximum parallel active scans (range: 1–10).
Max Payloads per Point	Number	10	Maximum payload variations per injection point (range: 1–50).
Timeout	Number	30	Timeout in seconds per scan request (range: 5–120).
Request Delay	Number	100	Delay in milliseconds between requests (range: 0–5000).
Max Risk Level	Dropdown	SAFE	Maximum payload risk level. Options: SAFE, MODERATE, DANGEROUS.
Scope Only	Toggle	On	Only scan in-scope targets.
Scan Mode	Dropdown	FULL	Vulnerability class selection strategy. Options: BUG_BOUNTY, PENTEST, FULL.
Auto-Queue from Passive	Toggle	On	Automatically queue high-confidence passive findings for active verification.
Use Collaborator (OAST)	Toggle	Off	Enable out-of-band testing with Burp Collaborator payloads.

Prompt Templates

Setting	Type	Description
Find Vulnerabilities	Text area	Default prompt for the "Find vulnerabilities" context menu action.
Analyze this request	Text area	Default prompt for endpoint summary.
Explain JS	Text area	Default prompt for JavaScript behavior analysis.
Access Control	Text area	Default prompt for authorization test plan generation.
Login Sequence	Text area	Default prompt for login flow extraction.
Analyze this Issue	Text area	Default prompt for scanner issue analysis.
Generate PoC & Validate	Text area	Default prompt for proof-of-concept generation.
Impact & Severity	Text area	Default prompt for impact and severity assessment.
Full Report	Text area	Default prompt for complete vulnerability report generation.

See Prompt Defaults for the full default text of each template.

File Locations

The extension stores data in ~/.burp-ai-agent/. All subdirectories are created automatically on startup.

File/Directory	Purpose
~/.burp-ai-agent/audit.jsonl	Audit log entries (JSONL format with SHA-256 hashing). Created when audit logging is enabled.
~/.burp-ai-agent/bundles/	Prompt bundle JSON files (used in determinism mode for reproducible analysis).
~/.burp-ai-agent/contexts/	Context snapshot JSON files (SHA-256 indexed).
~/.burp-ai-agent/certs/	Auto-generated TLS certificates (mcp-keystore.p12).
~/.burp-ai-agent/backends/	Drop-in backend JAR files. Place custom BackendFactory JARs here for ServiceLoader discovery. See Adding a Backend.
~/.burp-ai-agent/AGENTS/	Agent profile Markdown files. See Agent Profiles.
~/.burp-ai-agent/AGENTS/default	Text file containing the active profile name (e.g., pentester.md).

Runtime Defaults

Some operational limits are intentionally runtime constants (not direct UI fields):

• Active scanner queue max: 2000 targets (ACTIVE_SCAN_MAX_QUEUE_SIZE).

• Active dedup window: 1 hour (DEDUP_WINDOW_MS = 3,600,000).

• Passive analysis wait timeout: 90s (PASSIVE_SCAN_TIMEOUT_MS).

• HTTP backend conversation history cap: 20 messages (MAX_HISTORY_MESSAGES).

• CLI history cap: 10 messages or 20,000 characters (whichever is reached first).

• Large prompt threshold for Claude CLI fallback path: 32,000 characters.

Additional backend setting notes

• Ollama Timeout: Default 120 seconds (range 30–3600).

• Ollama Context Window: Default 8192 (range 2048–128000).