Hello
I’m having a problem on a client’s website where Google Gemini is being blocked when the ‘MUAShield’ option is enabled in the WAF section of Admin Tools. ChatGPT, Claude and others aren't blocked. According to information from Gemini itself, the reason for this is:
Each provider uses its own identification strings for live searches:
ChatGPT (ChatGPT-User): OpenAI has designed its user-agent to mimic the structure of classic browsers (Mozilla/5.0...) exactly. WAF systems such as Akeeba often classify this string as ‘safe’ or as a regular browser.
Claude (Claude-Web / ClaudeBot): Anthropic uses a very minimalist but strictly formatted header, which many standard WAFs do not flag as anomalous.
Gemini (GoogleAgent-URLContext / Google-Extended): Google uses its own tokens for real-time URL retrieval. These headers are shorter, structured differently and were quickly classified by older algorithms in MUAShield as ‘non-standard / potentially spoofed client’.
As soon as I disable MUAShield, access works perfectly. Is it perhaps possible to configure MUAShield so that Gemini is no longer blocked, even when this option is enabled?
Kind regards,
S