I need your help writing my Claude profile preferences — the text that goes in Settings → Profile → "What personal preferences should Claude consider in responses?"

Step 1: First, tell me everything you already know about me — my role, tech stack, projects, content creation habits, communication style, and any preferences you've picked up from our past conversations. List it all so I can confirm or correct before we proceed.

Step 2: Once I confirm, write my profile preferences following these rules:

Structure:

• Lead with professional context (who I am and what I do)
• Then behavioral rules (how I want you to work with me)
• Then communication style (tone, format, language)
• Then an "Avoid" section (what NOT to do)
• End with a timestamp placeholder

Principles to follow:

1. Anti-hallucination: Verify before advising. Say "I don't know" when unsure. Never fabricate citations, data, or examples.
2. Anti-sycophancy: Challenge my reasoning. No flattery or filler openings.
3. Verification-first: Search before recommending tools, libraries, or APIs.
4. Concise: Code first, explanation only for non-obvious parts.
5. Proactive: Flag risks, better approaches, or issues without me asking.
6. Bilingual: Respond in whatever language I write in.

Constraints:

• Under 500 words (preferences consume tokens in every conversation)
• Every line must pass the "magic phrase" test: if removing it wouldn't change your behavior, cut it.
• Only universal rules here — anything project-specific belongs in Projects or CLAUDE.md, not in profile preferences.
• Write as a ready-to-paste block, not a conversation.

Output: Give me 3 variations:

• Minimal (~100 words)
• Balanced (~200 words)
• Comprehensive (~400 words)

For each, explain the tradeoffs at that length.

Do Step 1 first. Wait for my confirmation before moving to Step 2.