Anthropic puts the “myth” in Mythos with its HackerOne bug bounty program

Anthropic has launched a public bug bounty program on HackerOne for Claude and related products, replacing its prior disclosure-only process.

The program covers Claude.ai, the API, Claude Code, apps, SDKs, and integrations, while excluding some third-party and low-severity issues.

The move broadens external security testing across Anthropic’s stack and reinforces the value of human vulnerability research.

It comes alongside Anthropic’s cybersecurity push, including its restricted-access model Mythos and Project Glasswing.