Skip to content

Latest commit

 

History

History
111 lines (84 loc) · 2.92 KB

File metadata and controls

111 lines (84 loc) · 2.92 KB

Agent Tools

Strix agents have access to a suite of specialized tools for security testing. These tools enable comprehensive vulnerability discovery and validation.

Available Tools

Tool Purpose
browser Multi-tab browser automation for XSS, CSRF, auth flow testing
proxy HTTP proxy for request/response interception and manipulation
terminal Interactive shell for command execution in sandbox
python Python runtime for exploit development and validation
file_edit Source code analysis and modification
web_search Information gathering via web search
notes Knowledge management and findings documentation
reporting Vulnerability report generation
agents_graph Multi-agent orchestration and coordination
thinking Internal reasoning and planning
finish Scan completion and result finalization

Tool Details

Browser

Playwright-based browser automation for testing client-side vulnerabilities:

  • Multi-tab support for complex attack flows
  • Screenshot capture for evidence
  • Cookie and session manipulation
  • JavaScript execution
  • Form interaction and submission

Requires a vision-capable LLM model for screenshot analysis.

Proxy

Full HTTP/HTTPS proxy for traffic interception:

  • Request/response modification
  • Header manipulation
  • Parameter tampering
  • Replay attacks
  • Traffic logging

Terminal

Sandboxed shell environment:

  • Command execution
  • Tool installation (curl, wget, etc.)
  • Network operations
  • File system access within sandbox

Python

Custom Python runtime:

  • Exploit development
  • Payload generation
  • Response parsing
  • Cryptographic operations
  • Custom validation scripts

File Edit

Source code analysis:

  • Read and analyze code files
  • Identify vulnerable patterns
  • Track data flows
  • Modify files for testing

Web Search

External information gathering (requires PERPLEXITY_API_KEY):

  • CVE research
  • Technology fingerprinting
  • Documentation lookup
  • Known vulnerability patterns

Notes

Internal knowledge management:

  • Store findings during scan
  • Track discovered endpoints
  • Document attack paths
  • Maintain context across agents

Reporting

Vulnerability documentation:

  • Generate structured findings
  • Include proof-of-concept details
  • Provide remediation guidance
  • Export to standard formats

Agents Graph

Multi-agent coordination:

  • Spawn specialized sub-agents
  • Coordinate parallel testing
  • Share findings between agents
  • Manage agent lifecycle

Sandbox vs Non-Sandbox Mode

Some tools behave differently based on execution mode:

Tool Sandbox Non-Sandbox
terminal Isolated container Host system (caution)
file_edit Container filesystem Actual target files
reporting Available Available
agents_graph Limited Full access

By default, Strix runs in sandbox mode using Docker for isolation.