Prompt-craft & iterating

🎯Goal. Get better, faster results by being specific, breaking work into steps, and letting Claude explore before it edits.

1. 1Be specific. Instead of "fix the bug", say "fix the parser bug where rows with a missing e-value are silently dropped". Specific requests let Claude work with minimal file reads.
2. 2Break complex tasks into numbered steps: 1) parse the BLAST output, 2) keep the top hit per query, 3) write a CSV and plot an e-value histogram. Claude follows the list in order.
3. 3Let Claude explore first — ask "analyze this analysis pipeline" before "now refactor it". If it heads the wrong way, press Escape to stop immediately and re-steer.

✓You'll see. The same task done two ways — a vague prompt that wanders and reads half the repo, and a specific one that lands the change in a couple of file reads.

💳Cost. Tighter prompts mean fewer tokens — vague asks like "improve this codebase" trigger broad scanning; specific asks keep usage (and your bill) down.

💡Takeaway. Treat Claude like a capable colleague: say exactly what "done" looks like, let it read first, and stop it early if it drifts.