Menu

Star on GitHub · 3.1k
All posts
News

Cyber Security News covers CyberStrike's automated pentest platform

Cyber Security News profiled CyberStrike as an AI-powered platform that automates reconnaissance, vulnerability discovery, exploitation and reporting from the terminal.

CyberStrike AI-powered penetration testing platform

Cyber Security News published a profile of CyberStrike, framing it as an AI-powered security platform built for automated penetration testing rather than another chatbot wrapper. Here is what the piece describes, in our own words, with the source linked at the end.

What the platform does

Per the article, CyberStrike installs as a terminal-based tool that runs the full assessment loop, reconnaissance, vulnerability discovery, exploitation and reporting, without a human prompting each step. The framing is that it turns an existing LLM subscription into an autonomous red-team operator, aimed at penetration testers, bug bounty hunters and security teams trying to scale their assessments.

The intelligence layer

The article’s main point is that the useful part is not the model, it is the layer around it. CyberStrike injects OWASP testing methodology, known vulnerability patterns and attack-chain reasoning into every model interaction, so the underlying model does not need built-in security knowledge. That layer is described as schema normalization, context guarding, provider auto-detection and tool orchestration.

Models and coverage

Cyber Security News reports support for 150+ AI providers and 5,300+ models, including Anthropic, OpenAI, Google, Amazon Bedrock and Azure, plus fully offline options like Ollama and LM Studio for air-gapped environments. It also cites 13 or more domain-specific agents: web application testing on OWASP WSTG, mobile on MASTG and MASVS, cloud around CIS benchmarks, and internal network testing covering Active Directory and lateral movement.

Our read

This is press coverage, not our own benchmark, so treat the capability claims as the outlet’s. What is accurate to state from our side: CyberStrike is an AI-augmented penetration testing project, open source under AGPL-3.0, and still pre-launch. The honest version is that automation shortens the repetitive parts of an assessment; a human still owns scope, judgment and sign-off.

The same tooling shows up in the FortiBleed reporting, where a criminal operator misused it. For that side of the story:

Source

Keep reading

More from on call

View all posts
Arctic Wolf Labs report: Inside FortiBleed, reverse-engineering the CyberStrike Harvester
newsJun 23, 2026

Arctic Wolf reverse-engineers the FortiBleed harvester

Arctic Wolf Labs tore down the Go-based credential tool recovered from FortiBleed, showing how perimeter captures on FortiGate firewalls turn into full internal-network compromise.

FortiBleed: a red FortiGate firewall leaking harvested credentials
newsJun 23, 2026

The Hacker News: FortiBleed hit FortiGate with no new zero-day

The Hacker News reported on FortiBleed, a Russian-speaking access broker that harvested 110M+ credentials from FortiGate and other gateways, without a single new vulnerability.

SoCRadar FortiBleed exposure checker
newsJun 20, 2026

SoCRadar's 'Dismantling FortiBleed' maps the full operation

SoCRadar's Threat Research Unit published a 77-page teardown of FortiBleed: 430,000+ FortiGate firewalls in scope, 659 harvest cycles, 110M+ credentials, and a breach at a NATO-aligned defense contractor.

Get the next post in your inbox

One email a month with new posts, product updates and the odd postmortem worth reading. Unsubscribe in one click.