# Meta details how Catching JiTTests generates bug-finding tests before code lands

> Original title: Catching Bugs Before They Land: JiTTesting’s Human-on-the-Loop Approach at Full Scale

- Source: Engineering at Meta
- Published: 2026-10-29T16:00:12.000Z
- Bad Influence: https://badinfluence.news/items/tgzw8fcj6gmmis2lz62nzh6m2
- Original: https://engineering.fb.com/2026/10/29/security/catching-bugs-before-they-land-jittestings-human-on-the-loop-approach-at-full-scale/

## Summary

Meta says its Catching Just-in-Time Tests (JiTTests) have moved from research prototype into production, where they generate executable tests that fail on code changes before they land. The system uses two generation workflows and filters weak catches with an LLM-as-Judge ensemble and a rule-based assessor, then has engineers review results at an aggregate level. Meta reports that 589 engineers engaged with it, with a 60% true-positive rate, and that it helped catch 74 serious privacy bugs in the first half of 2026, lowering privacy SEVs by 20%.

## Why it matters

The post explains how the method is built and reports deployment results, which help readers judge whether AI-generated tests can work at large codebase scale.