I read a post by sava.rocks [0] and various replies on the
geminispace BBS [1] about AI agents grabbing content from gopher and
gemini using an MCP server developed last year. It really was just a
matter of time before they infiltrated the smolnet.

The developer of the MCP server in question ('gopher-mcp'), has
changed the code to honour robots.txt, and added some rate limiting.
Better than nothing, I suppose.

Meanwhile, a user on the SDF bboard posted a sensible, anti-AI
robots.txt for gopher (it should work for gemini as well).

    User-agent: gopher-mcp
    Disallow: /

    User-agent: GPTBot
    Disallow: /

    User-agent: ChatGPT-User
    Disallow: /

    User-agent: OAI-SearchBot
    Disallow: /

    User-agent: ClaudeBot
    Disallow: /

    User-agent: anthropic-ai
    Disallow: /

    User-agent: Claude-Web
    Disallow: /

    User-agent: Claude-User
    Disallow: /

    User-agent: Google-Extended
    Disallow: /

I want to say a better choice is to disallow everything, and only
allow, say the veronica gopher search crawler (or any others you may
want to allow), but I don't know if veronica's crawler logic can
parse that type of "default deny but allow these few things" syntax,
and it may end up allowing too much if it doesn't.

And something fun, if you use GNU Emacs to write on the smolnet, you
can use 'M-x spook' to add some "subversive", random nonsense words
to each post, to give the agents or bots something to chew on. Here
is an example M-x spook output:

    Influenza MILSATCOM Critical infrastructure Rule Psix blackjack
    distributed denial of service FMD UDT 9/11 Denial of service
    AVIP VIP Protection Immigration Customs Enforcement Leitrim
    Airplane

[0]: gopher://sava.rocks/1/blog/when-the-small-web-gets-an-ai-door/
[1]: gemini://bbs.geminispace.org/u/SavaRocks/47932