The Hookbuilder now has an automated feedback loop:
An AI agent checks our Discord hourly, fixes code issues and turns useful ideas into improvements, so progress no longer depends only on our build speed.
We also plan to reward active contributors who find and report these bugs and ideas
https://t.co/FkaYhRbiwe
Hookbuilder v0.9 is live.
What changed:
-Understands natural @Uniswap v4 build requests without requiring special Programmable wording
-Loads specialist knowledge only after the project intent is confirmed
-Supports complete nested Foundry contract and test structures
-Diagnoses failed builds through a bounded repair flow instead of blindly retrying
-Produces much smaller project reports without removing the underlying evidence
-Uses the same central launch requirements as the submission system
Next, we’re building a benchmark system that runs the same real projects through different Hookbuilder versions and other builder setups.
https://t.co/55Q9uMssSU
We just shipped Hookbuilder v0.8.0.
For our upcoming test launches, we intentionally reduced the central launch policy to the minimum.
The bigger change is structural:
The policy is now the single source of truth. Hookbuilder builds against the same requirements our launch and security checks evaluate, and external agents can target the exact same contract.
That makes submissions much closer to one-shot:
fewer mismatches between what gets built and what gets reviewed.
Security findings can still block approval, but policy drift should no longer be the reason.
Hookbuilder: https://t.co/HDty43mXat
Policy:
https://t.co/MI9fxPdmra
Welcome to the technical side of Programmable.
This is where we share the work behind the product and go deeper into the details than we do on the main account.
Code: https://t.co/mU3wosX9lf
Docs: https://t.co/O97QiUJj8t
Community: https://t.co/IZ8BiJXJFJ