AI system prompts

Repo

→ CL4R1T4S → Tracks reverse-engineered system prompts → Covers ChatGPT, Gemini, Grok, Claude, Perplexity, Cursor, others

Why it matters

→ Knowing operating rules = better prompt design → Spot when AI is agreeable instead of accurate → Spot when AI stays polite instead of pushing back → Spot when AI fabricates an answer it doesn’t have WARNING: trusting tone over content = bad calls

Migrated from old article. Review for current relevance.