AI Wiki · Mediocre Security
ProductAgent products & apps

Opera Neon

Opera's agentic browser, central to its bet on bespoke browsers for the agentic web.

Opera’s agentic browser, central to its bet on bespoke browsers for the agentic web (PCMag, Oct 24 2026).

Security and assurance angle

Another browser where an agent acts with the user’s logged-in sessions.

Prompt-injection flaw

Brave’s researchers reported that hidden HTML elements on a web page could steer Neon’s AI into misusing the browser’s tools, a textbook indirect prompt injection. Brave’s timeline says Opera first closed the report on Oct 17 2025 as “Not Applicable”; Opera says the issue came in through its bug bounty and was addressed. Opera’s mitigations include analyzing prompts for malicious traits, restricting AI actions to a safe subset of browser capabilities, and separating the user’s prompt from untrusted site content, while conceding the risk can’t be reduced to zero. The UK NCSC has warned prompt injection may never be fully mitigated. Every agentic browser, including Strawberry and ChatGPT’s browser control, faces the same problem.

Timeline

Every dated event in the wiki that involves Opera Neon, newest first.

  • IncidentOpera first closed the report Oct 17 2025; later addressed
    Prompt-injection flaw in Opera Neon reported by Brave

    Hidden HTML on a page could steer Neon's AI into misusing browser tools, an indirect prompt injection. Opera says it fixed the issue and that the risk cannot be reduced to zero.

Connections

Opera Neon and everything one link away. Drag to move, scroll to zoom, click a node to open it.Open in the full graph →

What links here

Product Strawberry (browser agent)

Sources

  1. PCMag, Oct 24 2026 [link needed]
  2. Brave, “Prompt injection flaw in Opera Neon”
  3. Opera, “Opera Neon: understanding agentic browser security”, Oct 2025
  4. Gizmodo [link]