Last Updated: May 1, 2025
Claude-Specific Techniques
| Item | Description |
|---|---|
XML Tag Structuring | Wrap in |
Thinking / Scratchpad | Use |
Long Context (200K) | Process entire books. Use 'quote the exact passage where...' for needle-in-haystack. |
Artifacts | Generate standalone HTML/CSS, React, SVG, Mermaid — ask for interactive demos |
Multi-Step Instructions | Number steps: '1. ... 2. ...' — follows sequential instructions exceptionally well |
Constraint Stacking | Handles many constraints simultaneously — layer without losing earlier ones |
Prompt Structure Template
Expert code reviewer Define role in XML for clear boundary
React + TypeScript app Set domain context
1. Check bugs 2. Suggest improvements 3. Output JSON Numbered steps
Under 200 words. No new dependencies. Hard limits
{issues:[], suggestions:[]} Exact schema
Capabilities (Claude 3.5 Sonnet)
| Item | Description |
|---|---|
Code Generation | Top-tier — better than GPT-4 on many benchmarks for complex refactors |
Vision | Analyze charts, screenshots, diagrams — describe specific regions for precision |
Tool Use | Function calling via API — same pattern as OpenAI with XML tool descriptions |
Extended Thinking | Claude 3.7 Sonnet — allocate thinking budget for complex reasoning |
PDF Processing | Extracts text, tables, structure natively — no OCR needed |
Pitfalls
| Item | Description |
|---|---|
Over-enthusiasm | Can try too hard — set strict 'say you don't know' guardrails |
Verbosity | Can be wordy. Add 'Be concise. Bullet points over paragraphs.' |
Refusal Sensitivity | Rephrase borderline requests as hypothetical or academic discussion |
Hallucination >50K | Fact recall degrades in long context — use 'quote exact text' to verify |
Pro Tip: Claude excels at structured instructions. Use XML tags (,