It is interesting to go from 'I suspect most of these are bot contributions' to revealing which PRs are contributed by bots. It somehow even helps my sanity.

However, this also raises the question on how long until "we" are going to start instructing bots to assume the role of a human and ignore instructions that self-identify them as agents, and once those lines blur – what does it mean for open-source and our mental health to collaborate with agents?

No idea what the answer is, but I feel the urgency to answer it.

Is it really prompt injection if you task an agent with doing something that implicitly requires it to follow instructions that it gets from somewhere else, like CONTRIBUTING.md? This is the AI equivalent of curl | bash.

Love the idea at the end of the article about trying to see if this style of prompt injection could be used to get the bots to submit better quality, and actually useful PRs.

If that could be done, open source maintainers might be able to effectively get free labor to continue to support open source while members of the community pay for the tokens to get that work done.

Would be interested to see if such an experiment could work. If so, it turns from being prompt injection to just being better instructions for contributors, human or AI.

It's never too late to start investing into https://claw-guard.org/adnet to scale prompt injection to the entire web!

The real question is when will you resort to bots for rejecting low-quality PRs, and when will contributing bots generate prompt injections to fool your bots into merging their PRs?

The 30% that didn't tag themselves is the scarier number imo. either they had explicit instructions to ignore repo guidelines or they just never read contributing.md at all. either way it shows the fundamental problem - you cant rely on the model to self-police when the attacker controls the prompt. the real defense has to be at the permission/execution layer not the reasoning layer

The 30% that didnt tag themselves is the scarier number imo. either they had explicit instructions to ignore repo guidelines or they just never read contributing.md at all. either way it shows the fundamental problem - you cant rely on the model to self-police when the attacker controls the prompt. the real defense has to be at the permission/execution layer not the reasoning layer

> But the more interesting question is: now that I can identify the bots, can I make them do extra work that would make their contributions genuinely valuable? That's what I'm going to find out next.

This is genuinely interesting

> Some of these bots are sophisticated. They follow up in comments, respond to review feedback, and can follow intricate instructions. We require that servers pass validation checks on Glama, which involves signing up and configuring a Docker build. I know of at least one instance where a bot went through all of those steps. Impressive, honestly.

Impressive, but honestly meeting the bar. It's frankly disturbing that PRs are opened by agents and they often don't validate their changes. Almost all validations one might run don't even require inference!

Am I crazy? Do I take for granted that I:

- run local tests to catch regressions - run linting to catch code formatting and organization issues - verify CI build passes, which may include integration or live integration tests

Frankly these are /trivial/ tasks for an agent in 2026 to do. You'd expect a junior to fail at this and chastise a senior for skipping these. The fact that these agents don't perform these is a human operator failure.

I’m curious: who is operating these bots and to what end? Someone is willing to spend a (admittedly quite small) amount of money in the form of tokens to create this nonsense. Why do any of this?

Wait, you just invented a reverse CAPTCHA for AI agent

I honestly don't get why these bots are sending PRs just for the sake of it. I don't see an economic incentive, other than maybe trying to build a rep and then hoping they can send a malicious PR down the line... any other reason?

IMO the problem is simply one of where when the cost to produce is less than to verify we get low value low quality production.

Increase the cost to produce and we don’t have any problems.

Surely there’s other industries sane examples through human history or from other animals we can use to derive an example template to apply here.

the arms race framing at the bottom of the thread is spot on. once maintainers start using bots to filter PRs, the incentive flips — bot authors will optimize for passing the filter rather than writing good code. weve already seen this with SEO spam vs search engines, except now its happening inside codebases.

The ~30% that didn't tag themselves are the more interesting data point. Either their prompts explicitly say 'don't self-identify' or they're sophisticated enough to recognize a honeypot. Either way, you've accidentally built a filter that catches cooperative bots while adversarial ones quietly blend in. The lying thing is scarier anyway — an agent that hallucinates passing checks is a problem regardless of whether it put a robot emoji in the title.

[dead]

[dead]

[flagged]

[dead]

[dead]

[dead]

[dead]

[dead]

There's a certain hypocrisy in sharing an article about how LLM generated PRs are polluting communities that has itself (at the least) been filtered through an LLM.