HackerOne is integrating Anthropic’s Claude Mythos across select products in its H1 Platform, bringing frontier cyber AI into security workflows for vulnerability discovery, validation, prioritization, and remediation.
The integration will be available in H1 Code Security Audit and H1 Code. HackerOne said the technology is designed to help security teams validate AI identified vulnerabilities and bring confirmed exposures into existing engineering workflows.
According to H1 Platform data, critical vulnerability mean time to remediate improved by more than 50% over the past 12 months, while the unresolved critical backlog grew 29 times.
“Frontier cyber AI models can do one of two things to a security team,” said Nidhi Aggarwal, Chief Product Officer at HackerOne. “It can hand them a longer list, or it can hand them a list they can act on. Which one depends entirely on what sits between the model and the security team. A harness helps turn model output into a validated exposure a security team can act on. We ran and validated the H1 Platform harness for ourselves before any customer saw it.”
“Attackers are already using frontier AI to find and exploit vulnerabilities faster than humans alone can respond,” said Kara Sprague, Chief Executive Officer at HackerOne.“Defenders need access to advanced cyber-capable models, with people accountable for the decisions those models inform and the actions those models take. Finding vulnerabilities faster matters only if organizations can validate, prioritize, and act on them.”
HackerOne said frontier cyber AI will complement its security researcher community, which brings expertise across complex attack chains, business logic flaws, and novel exploitation paths. Select H1 Platform offerings featuring Mythos will be available soon.
