Secret Scanning and Credential Hygiene

Benchmarks are prototyping tools: how GitHub secret scanning ranked LLM configurations

Mariko, a Principal Applied Scientist at Microsoft, makes recall a gate rather than a tradeable metric. The structure travels well; the thresholds are not published.