Claude prompts engineered for long-context document review, careful reasoning, code refactors and tone-controlled writing.
Claude shines when you give it a lot of context and ask it to think carefully before answering. This pack leans into that: contract reviews, codebase walkthroughs, research synthesis, multi-document comparisons, editing in your voice, and structured extraction with confidence flags. Every prompt uses Anthropic-style XML tags and explicit output formats so Claude returns something you can hand to a teammate without cleanup.
Copy-and-paste starters for the most common claude jobs. No signup needed.
<task>Analyze the document below for [GOAL]. Think step by step before answering.</task> <document>[PASTE]</document> <output_format>1) Summary (≤80 words). 2) Key findings (bullets with quotes + page refs). 3) Risks. 4) Recommended next actions.</output_format>
Review my plan below and play devil's advocate. List the 5 strongest objections, the assumption each rests on, and how I'd know if the objection is right within 2 weeks. Plan: [PASTE]
The best prompts for claude give the model a clear role, the audience, the constraint, and the exact output format. Our Claude pack is structured that way — every prompt is a fill-in-the-blank you can paste into Claude or Claude and get a usable first draft in under a minute.
For most claude workflows, Claude is a strong default. Reach for Claude when you need long-context analysis or a more thoughtful, structured answer. Both work with the prompts in this collection — the scaffolds are model-agnostic.
Yes. The starter prompts on this page are free to copy, remix and use commercially. The deeper packs — with model-specific variants and 30–60 prompts per pack — unlock with a Pro plan.
Replace the bracketed placeholders with your specifics — audience, product, constraint. Keep the role and format instructions; that's what makes the output usable. If a prompt returns something generic, add one concrete example of a good answer and rerun.
Yes. Every starter here runs on the free tier of Claude. Paid tiers give better reasoning and longer context, but the prompts themselves don't require a paid plan.
Prompts for outlining, drafting, editing and adapting long-form writing without losing your voice.
Prompts for code review, refactors, debugging, architecture decisions and writing tests that actually run.
Proposals, SOWs, executive memos, frameworks and deliverables — the prompt library consultants wish they had on day one.
Pro unlocks the full library — every pack, every prompt, all models.
See pricing