Grisoco AI has opened its Canton security-review beta to developer feedback, with a tool built to examine Daml smart contracts.
The review focuses on four areas: who can take an action, who can see information, whether users compete for the same contract, and what an upgrade could change for contracts already running.
These checks address problems specific to how Canton applications are designed. A contract might give authority to the wrong party, expose details to someone who should not see them, or work well for one user but encounter conflicts when many people use it.
Grisoco's workspace uses several AI models to examine the code. Findings then go through additional review and challenge stages before being collected into one report. Developers can dispute a finding, submit a fix and have it checked again.
For permission-related findings, the tool can replay the scenario on a temporary Canton ledger to produce execution evidence. When developers supply the previous version of a package, it can also use the Daml compiler to check upgrade compatibility.
The beta is seeking feedback on missing checks, incorrect findings and useful integrations. Its results are automated, so developers should have a qualified engineer review them and arrange a professional audit before deploying contracts that hold significant value.
