Most Claude prompts for content creation<\/strong> you save will never get used a second time. They sit in a notes app, a Google Doc, a screenshot folder, and you rewrite them from memory anyway.<\/p>\n I know because I did exactly that for months. I had a folder of “great prompts” and I opened it perhaps twice.<\/p>\n A Claude prompt for content creation<\/strong> is a written instruction that tells Claude to produce or improve content, such as a blog outline, a social post or an email. The good ones name the audience, set the format and say what to avoid. They are for bloggers, newsletter writers and creators who write every week and are tired of generic output.<\/p>\n This post gives you 12 prompts worth keeping, grouped by the stage of work they help with. Then it shows you how to promote the ones you use again and again into reusable skills, so you stop pasting text at all.<\/p>\n A saved prompt fails when it has no context about you or your reader. Claude fills the gap with an average answer, and an average answer reads like everyone else’s.<\/p>\n Here is the thing. The prompt is only half the input. The other half is what Claude already knows about your voice, your audience and your rules before you type a word.<\/p>\n My first try at ChatGPT for writing felt flat, so I stopped. Many readers of this blog share that story. The tool was fine. The input was thin.<\/p>\n Does that mean you need a 600-word prompt every time? Not really. It means you need a short prompt that points at context Claude already has.<\/p>\n That is the idea behind everything below. Short prompts, with a place to store the context.<\/p>\n Every prompt below has square brackets where your details go. Replace them before you send.<\/p>\n Before you paste anything, do these three things once:<\/p>\n If you want the longer version of that setup, my post on building a Claude Project for your content voice<\/a> walks through it.<\/p>\n Planning prompts save you from writing the wrong thing. A bad topic costs you a full draft, so this is where a sharp prompt pays back most.<\/p>\n Use this before you commit to a topic. It stops you from writing something nobody is asking for.<\/p>\n Here is my reader: [paste your one-page description]. Here is a topic I am thinking of writing about: [topic]. Tell me honestly whether this reader would search for this, what question they are really asking, and what they would be disappointed to find. Do not flatter the idea.<\/p>\n<\/blockquote>\n The last sentence matters. Without it, Claude tends to agree with you.<\/p>\n Many topics already exist on your blog. This prompt finds the angle nobody has taken, including you.<\/p>\n My blog already has posts on these topics: [list titles]. I want to write about [topic]. Give me five angles that do not repeat what I have already covered, and for each one tell me which reader problem it solves.<\/p>\n<\/blockquote>\n Paste your real post titles. This is the prompt that keeps you from writing the same article twice by accident.<\/p>\n An outline is cheaper to fix than a draft. Ask for structure first and argue with it.<\/p>\n Build an outline for a blog post titled 12 Claude Prompts for Content Creation I Reuse Weekly for [reader]. Include a direct two-sentence answer at the top of each section, one section that is a numbered step-by-step, and one section on a common mistake and how to fix it. Keep it to eight sections or fewer.<\/p>\n<\/blockquote>\n Read the outline out loud. If a heading sounds like a stranger wrote it, change it before drafting.<\/p>\n Drafting prompts work best when you give Claude your raw material. Your stories and numbers are what make the piece sound like you.<\/p>\n A good hook beats a clever one. This prompt gives Claude the real moment and asks for the opening only.<\/p>\n Here is a real moment from my experience: [2 to 4 sentences, with the specific detail]. Write three possible openings for a post about [topic] that start inside that moment. Do not start with a definition. Keep each under 120 words and keep my plain speaking style.<\/p>\n<\/blockquote>\n Only paste a story that really happened. Claude will happily invent one if you let it, and readers can tell.<\/p>\n Drafting a whole post at once usually produces filler. Draft one section at a time.<\/p>\n Draft the section titled [heading] for a post aimed at [reader]. Start with a direct two-sentence answer. Use short paragraphs of three sentences or fewer. Include one specific number or named tool. Here are my notes: [paste notes].<\/p>\n<\/blockquote>\n Feeding it notes keeps the facts yours. You can run this eight times for eight sections and review each.<\/p>\n How-to content is what your reader came for. This prompt forces real mechanics instead of inspiration.<\/p>\n Turn these rough notes into a numbered step-by-step for [task]. Each step should be one action, start with a verb and say what the reader should see when it works. Add a “if this goes wrong” line under any step where people usually get stuck. Notes: [paste].<\/p>\n<\/blockquote>\n The “what the reader should see” line is the useful bit. It lets someone check their own progress.<\/p>\n Good posts answer pushback before the reader raises it. Ask Claude to play the sceptic.<\/p>\n Read this draft and list the three strongest objections a sceptical reader would raise. For each one, write a two-sentence reply I could add to the post in my own voice. Draft: [paste].<\/p>\n<\/blockquote>\n I use the replies as raw material and rewrite them. Taking them straight in makes the post sound defensive.<\/p>\n Editing is where Claude is most useful and least risky. It improves what you wrote instead of replacing it.<\/p>\n Most drafts carry 15 to 20 percent more words than they need. This prompt finds them.<\/p>\n Edit this draft to remove words that do not add meaning. Do not change my voice or my facts. Show me each cut as a before and after pair so I can reject any of them. Draft: [paste].<\/p>\n<\/blockquote>\n The before and after format is deliberate. You stay in control of every change.<\/p>\n AI-flavoured phrases creep in even with a good setup. Ask Claude to catch them.<\/p>\n Compare this draft to my voice rules below. Quote any sentence that sounds generic, corporate or like a template, and suggest a plainer version using my kind of wording. Voice rules: [paste]. Draft: [paste].<\/p>\n<\/blockquote>\n Run this last. If you want my longer routine for this step, read how I humanize AI content without losing the point<\/a>.<\/p>\n Do this after the draft is finished, not during. Writing for a keyword first makes everything stiff.<\/p>\n Check this post against the primary keyword and these secondary keywords [list]. Tell me where the primary keyword is missing from the title, first 15 words, headings and meta description. Suggest natural placements only. Do not stuff. Draft: [paste].<\/p>\n<\/blockquote>\n Natural placement is the rule. If a suggestion reads awkwardly, skip it.<\/p>\n One post can feed several channels if you do the conversion properly. These two prompts keep the meaning and change the shape.<\/p>\n A blog post and a Threads post need different rhythms. This prompt respects that.<\/p>\n Turn this blog post into a Threads thread of six posts. Each post should stand alone, stay under 450 characters and end with a reason to read the next one. No hashtags. Put “\u2013” on its own line between posts. Post: [paste].<\/p>\n<\/blockquote>\n If Threads is a main channel for you, my post on how to grow on Threads<\/a> shows what to put in those posts.<\/p>\n An email is not a summary of the post. It is a reason to click.<\/p>\n Write a short email that teases this blog post for my subscribers. Open with the problem in one sentence, give one useful point from the post, then link to the full piece. Under 150 words. Post: [paste].<\/p>\n<\/blockquote>\n Read it as a subscriber would. If it could be sent by anyone, rewrite the first line.<\/p>\n Here is how the prompts fit together on a single post. The order matters, because each step feeds the next one.<\/p>\n That is nine prompts for one post. It sounds like a lot. Each step is a short paste and a quick read, and the sequence replaces the blank-page hours at the start.<\/p>\n Bad output usually has a cause you can fix. Before you blame the tool, check these four things.<\/p>\n I made the first mistake for months. I kept asking for “a good post about X” and kept getting a post that could have been written for anyone.<\/p>\n Claude is a fast first-draft partner, but it has limits you should know about. Ignoring them is how people end up publishing text they cannot defend.<\/p>\n I have a simple rule. Claude can touch the shape of a post, but the facts and the stories stay mine.<\/p>\n A prompt you use every week should stop being a prompt. It should become a skill, which is a saved set of instructions Claude loads when it needs them.<\/p>\n Here is the difference. A prompt is something you paste each time. A skill is a named routine with its rules stored, so you call it by name and it runs the same way every time.<\/p>\n Look back over a month of chats. Any prompt you typed more than three times is a candidate. In my case the outline builder, the voice check and the post-to-thread converter qualify.<\/p>\n Write the steps in order, including what Claude should read first and what the finished output looks like. Add your voice rules and your banned words. This becomes the body of the skill.<\/p>\n Give it a plain name such as “outline-builder”. Run it on three different topics. If the output changes shape between runs, tighten the instructions until it stops.<\/p>\n Once you have a few, you can run them in sequence: plan, draft, edit, repurpose. That is where the weekly time saving shows up. I describe a full version of this in my post on automating a content workflow with Claude<\/a>.<\/p>\n Does every prompt deserve this treatment? No. A prompt you use once a quarter can stay a prompt.<\/p>\n Writing skills from scratch is slow, and the first few versions are usually rough. Mine were.<\/p>\n Want the routines already built and tested? The Content Creator’s Claude Skill Stack<\/a> is the set I put together for this work: planning, drafting, editing and repurposing. It also gives you working examples to learn from if you would rather build your own later.<\/p>\nWhy most saved prompts fail<\/h2>\n
How to use these prompts<\/h2>\n
\n
Prompts for planning<\/h2>\n
1. The reader check<\/h3>\n
\n
2. The angle splitter<\/h3>\n
\n
3. The outline builder<\/h3>\n
\n
Prompts for drafting<\/h2>\n
4. The story-first opener<\/h3>\n
\n
5. The section drafter<\/h3>\n
\n
6. The step-by-step writer<\/h3>\n
\n
7. The honest objection<\/h3>\n
\n
Prompts for editing<\/h2>\n
8. The cut-it-down edit<\/h3>\n
\n
9. The voice check<\/h3>\n
\n
10. The SEO pass<\/h3>\n
\n
Prompts for repurposing<\/h2>\n
11. The post-to-thread converter<\/h3>\n
\n
12. The email teaser<\/h3>\n
\n
What a full run looks like<\/h2>\n
\n
When the output is bad<\/h2>\n
\n
What to watch out for<\/h2>\n
\n
How to turn a prompt into a reusable skill<\/h2>\n
Step 1: Spot the prompts you repeat<\/h3>\n
Step 2: Write down the full routine<\/h3>\n
Step 3: Name it and test it<\/h3>\n
Step 4: Chain the ones that belong together<\/h3>\n
A shortcut, if you would rather not build it all<\/h2>\n