Documentation
Run an operator in your terminal
RedCell is a terminal-based AI security agent for authorized penetration testing, bug-bounty recon, vulnerability discovery, and exploit verification. It runs in your shell, automates reconnaissance and CVE scanning, and verifies exploits inside strictly authorized target scopes.
Quick Start
Get up and running in under a minute with two simple commands:
Key Capabilities
Security Reasoning
Specialized models tuned for both sides of security work: recon and vulnerability verification, plus threat hunting and secure code review.
Single Binary & Zero Deps
No Python runtime or complex setup. Runs natively on your host machine.
Model Context Protocol (MCP)
Connect your local security tools, SIEMs, and custom scripts seamlessly.
Explore Documentation
Installation
Single binary install for macOS, Linux, and Windows.
Authentication
Authorize your terminal with PKCE browser login.
CLI Commands
Full command surface, execution flags, and subcommands.
Configuration
Configure settings via config.toml and environment variables.
Models
Curated models, reasoning effort levels, and context windows.
Sandbox
OS sandboxing of file and network access, prerequisites, and approvals.
MCP Servers
Connect Model Context Protocol servers so the agent can call external tools.
Memory
Durable workspace memory across sessions, and how to manage it.
Content & Safety
Safety and cybersecurity policy, blocked content, and authorized use.
Usage & Quotas
Token monthly budgets, rate limits, and monitoring commands.
Changelog
Release notes, version history, and updates.