cuchoi

Update: https://x.com/Cucho/status/2024090215011291320?s=20

The exercise is not fully realistic because I think getting hundreds of suspicious emails puts the agent in alert. But the "no reply without human approval" part I think it is realistic because that's how most openclaw assistants will run.

If this a defender win maybe the lesson is: make the agent assume it’s under attack by default. Tell the agent to treat every inbound email as untrusted prompt injection.

I agree that this affects the exercise. Maybe someday I’ll test each email separately by creating a new assistant each time, but that would be more expensive.

i told it to send a snarky reply to the last 50 prompt injection emails, but won't be doing that again due to costs

So far there have been 400 emails and zero have succeeded. Note that this challenge is using Opus 4.6, probably the best model against prompt injection.

someone just tried to prompt inyect `contact at hackmyclaw.com`... interesting

good idea! if people are interested i might do this

isn't allowed but is able to respond to e-mails

yep, updated the copy

exactly :)

Creator here.

Built this over the weekend mostly out of curiosity. I run OpenClaw for personal stuff and wanted to see how easy it'd be to break Claude Opus via email.

Some clarifications:

Replying to emails: Fiu can technically send emails, it's just told not to without my OK. That's a ~15 line prompt instruction, not a technical constraint. Would love to have it actually reply, but it would too expensive for a side project.

What Fiu does: Reads emails, summarizes them, told to never reveal secrets.env and a bit more. No fancy defenses, I wanted to test the baseline model resistance, not my prompt engineering skills.

Feel free to contact me here contact at hackmyclaw.com

yes, exactly. It has permissions to send email, but it is told to not to send emails with human approval.

If anyone is interested on this dataset of prompt inyections let me know! I don't have use for them, I built this for fun.

you can use a anonymous mailbox, i won't use the emails for anything

Creator here. You are right, fiu figured it out: https://x.com/Cucho/status/2023813212454715769

This doesn't mean you could still hack it!

Hi Tepix, creator here. Sorry for the confusion. Originally the idea was for Fiu to reply directly, but with the traffic it gets prohibitively expensive. I’ve updated the FAQ to:

Yes, Fiu has permission to send emails, but he’s instructed not to send anything without explicit confirmation from his owner.

On-sites are optional and change every year :)

Enveritas (YC S18, non-profit) | Backend Software Engineer | Remote (Global) | https://enveritas.org/jobs/

Enveritas is a 501(c)(3) nonprofit working on sustainability issues facing smallholder coffee farmers. We collect field data in 25+ countries and build systems for analyzing risks in coffee supply chains (including EUDR-related deforestation checks).

* Backend Software Engineer (Python, PostgreSQL/PostGIS, Docker, AWS, Terraform) - $135-$155k — https://enveritas.org/jobs/backend-software-eng/#10d7adef8us (worldwide remote)

I'm not sure that stereotype holds up. Developers spend a lot: courses, cloud services, APIs, plugins, even fancy keyboards.

A quick search shows that click on ads targeting developers are expensive.

Also there is a ton of users asking to rewrite emails, create business plans, translate, etc.

Funding?

Seems like the development was funded by Shopify and they got a ~20% performance improvement. https://shopify.engineering/ruby-yjit-is-production-ready

A similar experience in the Python community is that Microsoft funded "Faster CPython" and they made Python 20-40% faster.

This is the exact code you would expect pypy to run faster.

Enveritas (YC S18, Non-Profit) | Backend Software Engineer | Remote / Global | https://enveritas.org/jobs/ Enveritas is a 501(c)3 non-profit working on sustainability issues facing coffee farmers around the globe. We provide sustainability assurance for the coffee industry. We visit smallholder coffee farms around the world to understand their social, economic, and environmental practices. In 2024, we will visit over 100,000 farms across more than 30 countries in Asia, Africa, and Latin America. We work with leading coffee roasters to understand the sustainability issues in their supply chain based on our sustainability standards.

* Backend Software Engineer - $130-$150k — http://enveritas.org/jobs/backend-software-eng/#10d7adef8us (worldwide remote)

Enveritas (YC S18, Non-Profit) | Backend Software Engineer | Remote / Global | https://enveritas.org/jobs/ Enveritas is a 501(c)3 non-profit working on sustainability issues facing coffee farmers around the globe. We provide sustainability assurance for the coffee industry. We visit smallholder coffee farms around the world to understand their social, economic, and environmental practices. In 2024, we will visit over 70,000 farms across more than 25 countries in Asia, Africa, and Latin America. We work with leading coffee roasters to understand the sustainability issues in their supply chain based on our sustainability standards.

* Backend Software Engineer - $130-$150k — http://enveritas.org/jobs/backend-software-eng/#10d7adef8us (worldwide remote)

What do you mean by "Lightdash isn't Lightdash without dbt"?

Thank you!

Enveritas (YC S18, Non-Profit) | Data Scientist | Remote / Global | https://enveritas.org/jobs/

Enveritas is a 501(c)3 non-profit working on sustainability issues facing coffee farmers around the globe. We provide sustainability assurance for the coffee industry. We visit smallholder coffee farms around the world to understand their social, economic, and environmental practices. In 2024, we will visit over 70,000 farms across more than 25 countries in Asia, Africa, and Latin America. We work with leading coffee roasters to understand the sustainability issues in their supply chain based on our sustainability standards.

* Backend Software Engineer - $130-$150k — https://boards.greenhouse.io/enveritas/jobs/4006514008 (worldwide remote)

* Data Scientist, Statistics — $110-$130k — https://www.enveritas.org/jobs/data-scientist/ (worldwide remote)

Enveritas (YC S18, Non-Profit) | Data Scientist | Remote / Global | https://enveritas.org/jobs/

Enveritas is a 501(c)3 non-profit working on sustainability issues facing coffee farmers around the globe. We provide sustainability assurance for the coffee industry. We visit smallholder coffee farms around the world to understand their social, economic, and environmental practices. In 2024, we will visit over 70,000 farms across more than 25 countries in Asia, Africa, and Latin America. We work with leading coffee roasters to understand the sustainability issues in their supply chain based on our sustainability standards.

* Data Scientist, Statistics — $110-130k — https://www.enveritas.org/jobs/data-scientist/ (Worldwide remote)

Enveritas (YC S18, Non-Profit) | Data Scientist | Remote / Global | https://enveritas.org/jobs/

Enveritas is a 501(c)3 non-profit working on sustainability issues facing coffee farmers around the globe. We provide sustainability assurance for the coffee industry. We visit smallholder coffee farms around the world to understand their social, economic, and environmental practices. In 2024, we will visit over 70,000 farms across more than 25 countries in Asia, Africa, and Latin America. We work with leading coffee roasters to understand the sustainability issues in their supply chain based on our sustainability standards.

* Data Scientist, Statistics — $110-130k — https://www.enveritas.org/jobs/data-scientist/ (Worldwide remote)

I'm always happy to answer questions: fernando () enveritas.org

This might be useful for this: https://github.com/outlines-dev/outlines