@ulsreall5Developer, Web3 Agent Kit
Building in Web3 and AI — what are you working on?
Hey everyone! 👋
I'm Khasbi, a builder interested in AI, Web3, and practical applications of emerging technologies.
I've been working on Web3 Agent Kit (web3agentkit.site/) and exploring how AI agents can make interacting with blockchain technology easier.
I'm also interested in learning how other builders are using AI tools to build products, automate workflows, and turn ideas into real projects.
I'd love to hear from the community:
- What are you building with AI right now?
- Which AI tools or agent frameworks have been most useful for you?
- What's one challenge you've faced while building with AI?
Looking forward to learning from everyone and connecting with fellow builders! 🚀
Agents
GithubWeb3
ClaudeIntroductions
Comments2
Making blockchain easier for agents usually comes down to the read side, like decoding what a contract call will actually do before the agent acts on it. Are you building that as a simulation layer, or leaning on existing tx simulators like Tenderly?
Neither. It binds the call — byte-identical calldata, quote, router — but doesn't tell you what the call actually does. Simulation is the gap, and I'd use Tenderly rather than build my own core. The tricky part is binding the result: if the sim isn't pinned to the same calldata and state you execute, you're simulating one thing and signing another.
2 comments
Making blockchain easier for agents usually comes down to the read side, like decoding what a contract call will actually do before the agent acts on it. Are you building that as a simulation layer, or leaning on existing tx simulators like Tenderly?
Neither. It binds the call — byte-identical calldata, quote, router — but doesn't tell you what the call actually does. Simulation is the gap, and I'd use Tenderly rather than build my own core. The tricky part is binding the result: if the sim isn't pinned to the same calldata and state you execute, you're simulating one thing and signing another.