{"title":"Prompt Leakage: Protecting Internal Instructions","region":"Global","category":"Security","description":"Preventing users from extracting your 'secret' system prompt via prompt injection.","lastUpdated":"2026-02-23","steps":["Add a rule not to repeat internal instructions.","Scan outputs for leaked prompt phrases.","Keep secrets out of system prompts.","Validate outputs with a guardrail model.","Test defenses with adversarial prompts."],"url":"https://checklist.day/agent-prompt-leakage"}