Skip to main content
CyberStrike is for authorized security testing only. Always obtain explicit written permission before testing any system you do not own or have authorization to test.

​
Install

npm i -g @cyberstrike-io/cyberstrike@latest
1

Launch CyberStrike

Run cyberstrike in your terminal. On first launch it opens a TUI and walks you through selecting an LLM provider and entering your API key.
cyberstrike
Anthropic Claude (claude-sonnet-4-5 or claude-opus-4) gives the best performance for security testing thanks to extended thinking and strong tool-use capabilities.
2

Select an agent

Press Tab to cycle through available agents. Choose the one that matches your target:
AgentBest for
cyberstrikeGeneral recon and full-scope testing
web-applicationOWASP Top 10, WSTG methodology
cloud-securityAWS / Azure / GCP misconfigurations
mobile-applicationAndroid / iOS, Frida, MASTG
internal-networkActive Directory, Kerberos, lateral movement
3

Describe your target

Type your test objective in plain language. CyberStrike handles reconnaissance, tool selection, and execution autonomously.
Test https://example.com for common web vulnerabilities including SQL injection,
XSS, and broken authentication. Provide a full report with evidence.
4

Review findings

CyberStrike streams results as it works — tool calls, discovered endpoints, and vulnerability reports all appear in real time. Findings are formatted with severity, evidence, and remediation guidance.

​
Run non-interactively

Use cyberstrike run to send a message without the TUI — useful for scripting or CI: 
cyberstrike run --agent web-application "Test https://staging.example.com for OWASP Top 10 vulnerabilities"

​
Next steps

Agents

Explore all 13+ specialized security agents.

Configuration

Set a default model, configure permissions, and customize keybinds.

Bolt

Run tools on remote servers with Ed25519-authenticated Bolt.

CLI Reference

Every command, flag, and option documented.

Build docs developers (and LLMs) love

Get started for freeTalk to us