Security Agents & Tools
A specialized fleet of AI agents working in concert — each an expert in its domain — to achieve the level of coverage and precision traditional tools cannot match.
Core Infrastructure Agents
Our agents are the foundation of our analysis pipeline, providing a comprehensive and accurate basis for every scan.
Vulnerability & Analysis Agents
These agents use offensive security techniques and semantic analysis of your application's source code and runtime logic to identify vulnerabilities.
Observation & Auditing Agents
For trust and security, it’s vital that every action taken by the AI can be audited and that it’s transparent. The following agents ensure that this requirement is met.
Tools Arsenal
Every agent has access to a purpose-built set of tools — from raw TCP sockets to AI-powered code delegation — giving the system full control over the attack surface.
File System
5 tools
Lists the contents of a project directory
Reads a full file with line numbers
Reads a specific line range from a file
Recursively searches files by name or glob pattern
Searches text within file contents
HTTP
4 tools
Performs a standard HTTP request (GET, POST, etc.)
Executes multiple requests in batch, in series or parallel
Iterates automatic payloads against an endpoint with success criteria
Sends raw bytes over TCP/TLS for full protocol control
Encoding
2 tools
Encodes text in base64, hex, URL, AES, RSA, JSFuck, PHP Filter Chain, etc.
Decodes text from base64, hex, URL, AES, RSA, etc.
Serialization
1 tool
Generates insecure deserialization payloads in Java, PHP, Phar or Python/Pickle
Semgrep
1 tool
Generates a Semgrep rule with AI and runs it against the project
Endpoints
3 tools
Registers a discovered API endpoint in the database
Registers multiple endpoints in batch efficiently
Detects endpoints with Semgrep and verifies them with AI before saving
Vulnerabilities
2 tools
Reports a vulnerability and starts the deduplication & verification pipeline
Searches for a related vulnerability in DB or code for chained exploits
Sub-Agents
3 tools
Launches a sub-agent that verifies the vulnerability by executing real attacks
Delegates a complex HTTP task to a specialised HTTP agent
Delegates a code question to a static analysis specialist agent