cuchoi

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 30+ countries and build systems for analyzing risks in coffee, cocoa, and tea 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)

There is a couple of factors: openclaw's system prompt and instructions, I had to re read emails multiple times due to the issues mentioned in the blog, there was quite a bit of tinkering with the agent and the VPS, I was asking the agent to do more things (track the emails it has read in a csv file, for example), among others.

The agent had permissions to reply to emails, it was just instructed not to.

You need to add Openclaw's system prompt and instructions (and the times I had to re read emails multiple times due to multiple issues that happened during the competition :))

The agent did read the emails

We ended increasing the reward from $100 to $1000, but still tiny compared to $100k!

But I agree with you, there are incentives to not share the best prompt injection attacks.

I never set out to spend this amount! Was able to keep it up thanks to the sponsors that reached out.

This openclaw was set up exclusively for the challenge.

100%. I am less worried because I thought this would be easier to crack.

In my case, it is realistic as my agents don't have permissions to reply to emails. But you correctly point out this doesn't cover all cases.

Having the agent reply would have been more fun and a better excercise, but too expensive.

Did you send this recently? I turned off the agent. Was too expensive to keep it up.

Author here, that's how I meant it. I changed my mind slightly, prompt injection can still happen, I am still careful.

Agreed. I am less worried about prompt injection now, but I still haven't given my agents permissions to send emails.

I changed the setup so that each email was processed in a fresh context. For this, I deleted recent memory and processed each email one at a time. Edited the post to make it more clear.

Thanks for sharing your article, very interesting.

I used https://github.com/openclaw/openclaw-ansible and configured a heartbeat (using Openclaw's terms) to check emails every hour. Had to do a bit more to make sure it had new context for every email.

About 1), Google didn't remove a lot of the attempts. I had also Fiu review the Spam folder as well.

Also, I mentioned how I addressed 2) by having new context for each email.

It's possible. I implemented something similar when I figured out that batch processing contaminated the excercise.

Author here. It was usable like any Openclaw agent. For example, I used it to ask it questions about the VPS, to summarize emails, etc.

Author here. Edited the post to clarify that there were no unauthorized replies.

I did tell Fiu initially to reply to some emails as a test, but it was too expensive to maintain.

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)

Hi /u/munimdev. We prefer for all applications to apply through the link above. This way the process stays fair for everybody. If you are on the fence about applying, send me you resume to fernando at enveritas org.

We're a small team and every hire makes a large impact, so we're careful and thorough when it comes to hiring, but we are actively considering candidates and moving them through the process as quickly as we can.

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)

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)

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)

yes, exactly

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.