Opus 5.5 Finds SDK Bugs via Lean Formal Verification
Formal verification is emerging as a practical bug-hunting workflow rather than an academic exercise, and a frontier model writing Lean and TLA+ proofs lowers the barrier for teams without verification expertise. Engineering leads working on concurrent or stateful systems can replicate this prompt pattern against their own codebases. The 16-PR yield from a couple of prompts gives a concrete sense of the payoff.
Boris Cherny used Claude Opus 5.5 to formally verify the Claude Agent SDK with the Lean proof assistant.
A couple of short prompts produced 16 pull requests fixing bugs and race conditions.He also combined Lean with TLA+ to probe data flow, concurrency, and state management issues, noting Claude performs well in both languages even for users who don't know them.
If you’re looking for ways to try Opus 5.5 and Claude Tag in Slack… Boris Cherny @bcherny I used Opus 5.5 to formally verify the Claude Agent SDK using Lean. A couple short prompts = 16 PRs fixing various bugs and race conditions. Video attached. TLA+ also works well. I sometimes combine Lean and TLA+ to look for issues around data flow, concurrency, and state mgmt. I don't know either language well, but Claude is excellent at both. This approach is super useful for formally modeling your code and finding bugs that a human probably wouldn't have spotted. Is formal verification the future of coding (or at least, bug finding)?
This report is categorized under AITOP's "How-to" desk, focusing on US frontier AI labs and Silicon Valley technology leaders. Within our cross-regional analytical framework, it directly maps to AI developer toolchains, coding agents, and engineering workflows. Key entities and technical topics under continuous tracking include Opus 5.5, Claude Agent SDK, Lean, TLA+.
Source attribution: curated from cat (@_catwu).
