Baret

Pre-sign checks & passkey-native wallets for every AI agent. | Know before you sign.

Built on @monad_xyz , @StellarOrg , @Casper_Network

[Baret @BuildOnBaret]

Check in Chrome : https://t.co/XGz2DZRQ3J

[Baret @BuildOnBaret]

Baret is now live on Chrome!

Before you sign a Monad transaction, Baret simulates it, checks it against 25 customizable security rules, and shows you exactly what you're approving.

Safe, Caution, or Blocked.

You'll know why before you sign.

Don't sign blind.

[Baret @BuildOnBaret]

One signature can do more than you think.

Baret is coming to @monad

Simulate before signing.
Detect risky approvals.
Enforce on-chain spending caps for AI agents.

Security before you sign.

Built for Metropolis.

[Baret @BuildOnBaret]

@ClaudesRH 👀

[Baret @BuildOnBaret]

If my wallet could do anything?

I’d make it understand the transaction before I sign it.

Simulate. Detect malicious behavior.

Monitor approvals. Revoke dangerous allowances.

Warn me before I sign.

Basically, Baret. 👷🏻‍♂️

[MetaMask 🦊 @MetaMask]

if your AI agent wallet could do anything, what would you have it do?

[Baret @BuildOnBaret]

The future of wallet security isn't just about protecting users from malicious transactions.

It's about controlling what autonomous agents are allowed to do with your assets.

Baret simulates transactions before signing, detects malicious behavior, and brings enforceable spending limits to agent-driven payments.

That's the problem we're building for.

Thank you @Casper_Network for recognizing Baret. 🛡️

[Casper @Casper_Network]

Letting agents pay for things was always going to create problems.

Baret is a Casper wallet that simulates every transaction before you sign it.

To show what Casper catches, they built six fake dApps with real attacks inside.

Five are scams you'd recognise. The sixth is an agent spending a little more each time, which nobody was worrying about five years ago.

Let's dive into one of the finalists in the Casper Agentic Buildathon.

The six traps

1. NovaSwap. A DEX clone. A hidden instruction sends your tokens to a stranger.

2. PixelDrop. An NFT mint that quietly drains every token you hold.

3. OrbitYield. Liquid staking at 14%, with no unstake path. Money goes in only.

4. ClaimHub. An airdrop claim whose "eligibility check" signs away your stablecoins.

5. LaunchPad. A token launch where the dev kept mint authority and the liquidity isn't locked.

6. Scrybe. An AI service charging per answer over x402, testing whether your wallet caps an agent's spend.

What happens when you press sign

- A server simulates the unsigned transaction

- 25+ detectors run against it

- You get one sentence explaining what's wrong

- You sign with your eyes open, or you reject

Why the last one matters

The first five have someone trying to rob you. The sixth doesn't.

Scrybe is a legitimate service charging a fair price, and the agent paying it is doing exactly what it was told.

Which is why the limit has to be a rule the wallet enforces, rather than a judgement someone makes in the moment.

[Aeztrest @Aeztrest]

This is exactly why I built Baret.

Wallet security can't stop at detecting scams.

As AI agents start making transactions on our behalf, we need wallets that can understand, simulate and enforce what an agent is allowed to do.

Baret is built for that layer.

[Casper @Casper_Network]

Letting agents pay for things was always going to create problems.

Baret is a Casper wallet that simulates every transaction before you sign it.

To show what Casper catches, they built six fake dApps with real attacks inside.

Five are scams you'd recognise. The sixth is an agent spending a little more each time, which nobody was worrying about five years ago.

Let's dive into one of the finalists in the Casper Agentic Buildathon.

The six traps

1. NovaSwap. A DEX clone. A hidden instruction sends your tokens to a stranger.

2. PixelDrop. An NFT mint that quietly drains every token you hold.

3. OrbitYield. Liquid staking at 14%, with no unstake path. Money goes in only.

4. ClaimHub. An airdrop claim whose "eligibility check" signs away your stablecoins.

5. LaunchPad. A token launch where the dev kept mint authority and the liquidity isn't locked.

6. Scrybe. An AI service charging per answer over x402, testing whether your wallet caps an agent's spend.

What happens when you press sign

- A server simulates the unsigned transaction

- 25+ detectors run against it

- You get one sentence explaining what's wrong

- You sign with your eyes open, or you reject

Why the last one matters

The first five have someone trying to rob you. The sixth doesn't.

Scrybe is a legitimate service charging a fair price, and the agent paying it is doing exactly what it was told.

Which is why the limit has to be a rule the wallet enforces, rather than a judgement someone makes in the moment.

[Baret @BuildOnBaret]

The next billion crypto users won't lose money because blockchains got faster.

They'll stay because wallets finally become safer.

That's what we're building with Baret on Stellar.

🛡️ Prevent mistakes before they happen.

#Stellar #Web3

[Baret @BuildOnBaret]

Crypto isn't hard. Staying safe is.

Baret helps users detect risky transactions before they sign them on Stellar.

No more blind approvals.
No more wallet-draining scams.
Security should be the default, not an add-on.

GitHub: https://t.co/jvKHUmW1l1

#Stellar #Soroban #Web3 #Blockchain

[Baret @BuildOnBaret]

The real problem with AI agents is not intelligence.

It is authority.

Baret does not trust the model with your wallet.

It decodes every transaction, applies spend limits and policy checks, and only then allows signing.

AI can suggest.
Rules decide.
Humans stay in control.

GitHub: https://t.co/1lbnkLsVCn

[Build on Stellar @BuildOnStellar]

well would you look at the time...

12
11 ^ 1
10 | 2 it’s Build On
9 ⊙———> Stellar
8 4 o’clock
7 5
6

[Build on Stellar @BuildOnStellar]

。
ｏ
○
╭◜◝ ͡ ◜ ͡ ╮ ╭◜ ◝ ͡ ◜◝ ͡ ╮
( build ) ( on )
╰◟◞ ͜ ╭◜◝ ͡ ◜ ◝ ͡ ◝ ͡ ╮◞◟◞ ╯
( stellar )
╰◟◞. ͜ ◟◞ ͜. ◟◞ ╯