Formal Verification Beats Code Review: Catches 6/8 OSS Bugs vs Just 1/8
StefanoGogioso · x · 2026-10-25
Doug (@0xdoug) explains how his team achieved exceptionally high coverage on major real-world bugs in deeply reviewed codebases: the key is formal verification, and how you prompt the agent to draft the formal spec matters a lot.
- Code review alone would have caught roughly half of the crypto exploits and a paltry 1/8 of the OSS exploits.
- Formal verification would likely have caught all of the crypto exploits and 6/8 of the OSS exploits.
The takeaway: instead of having agents merely read code to hunt bugs, guide them to formalize the code's intent as a spec and verify it — coverage improves dramatically. The author says they're sharing the exact prompting approach.
More from coding & agent
- Jev ecosystem explodes: 5 open-source apps from a 7.1s flight-search agent to SQL-in-English — udmrzn · 2026-10-25
- PhD student says Claude now does most of his thesis work, down to the draft — GarrisonLovely · 2026-10-25
- Google's Jon Barron shares his full AI paper-writing workflow with Gemini — lmoroney · 2026-10-25
- Dev runs an Instagram agent on night shift: 145 accounts checked, 10 follows — petrusenko_max · 2026-10-25
- Minga lets you manage coding agents at the level of goals and plans — every · 2026-10-25
- 1Password lets AI agents sign into your accounts without ever seeing your passwords — LuizaJarovsky · 2026-10-25