Project summary

IDRAIN is a free public tool that checks the system prompt of an AI agent and tells you how easy it is to trick that agent into sending real money.

It looks for missing basic protections — spending limits, human confirmation, allowlists — and shows if the prompt is vulnerable to the same kind of prompt injection attacks that already stole money in 2026 (including the Grok/Bankrbot case).

Demo is live and the code is open source. Right now the only real problem is keeping the tool online and slowly improving it.

What are this project's goals? How will you achieve them?

I just want it to be stupidly easy for people to check their agent prompts before they give the agent access to money.

How I’ll do it:

• Keep the tool free and fast

• Update the detection rules when new real attacks appear

• Make the interface simple enough that even non-security people can use it in 20–30 seconds

• Leave everything open source so others can improve it if they want

How will this funding be used?

Mostly just hosting (Vercel + domain) for the next 12 months. A bit of money for small improvements and basic analytics. No salary, no team, no fancy stuff. Just keep the tool alive and free.

Who is on your team? What's your track record on similar projects?

Just me. I work on the intersection of AI agents and crypto/money. I built this MVP in one evening because I got tired of seeing the same stupid mistakes in agent prompts. I prefer shipping small useful tools over writing long papers.

What are the most likely causes and outcomes if this project fails?

Most likely:

• Almost nobody uses it

• Someone makes a better version

If it fails — the code stays open source, people can still fork it, and only a small amount of money for hosting will be spent. Pretty low downside.

How much money have you raised in the last 12 months, and from where?

$0