CONCEPT
Leak Detection Testing
Testing regimes designed specifically to find the places where AI-generated code is most likely to fail — concurrency, integration boundaries, failure injection, current-threat security scanning — before production conditions force the discovery under time pressure.
Leak detection testing is this volume's prescriptive extension of Spolsky's framework: a family of testing practices targeted not at validating specified behavior (what standard tests do) but at discovering behaviors the specification did not address, which is where leaks live. The practice has four principal categories, each targeting a specific leak class: concurrency testing for race conditions, integration boundary testing for assumption mismatches between generated components, failure injection for cascading failures under degraded conditions, and security scanning against current threat intelligence (not just the threats encoded in training data). The practices are not exotic — each is well-understood in contexts where systems must not fail — and their novelty lies in the systematic application to AI-generated code, which does not apply them to itself.
In The You On AI Field Guide
Standard testing validates that a system does what the specification says it should do. Unit tests check that functions return expected outputs. Integration tests check