This is the actual prompt chain I run every morning. Five minutes of capture in CTRL App goes through Claude and comes back as a full day of content across Instagram, with no team, no agency and no VA behind the scenes.
Five pieces of content a day go out across Instagram. Reels, carousels and captions, all of it run by one person with no team, no content agency and no VA scheduling posts in the background. The whole pipeline runs through AI and it starts the same way every morning, with a raw journal entry that has not been tidied up or pre-planned.
The real unlock is not a clever prompt. It is the system sitting underneath it, a set of custom skill files I have built inside Claude that already know the brand voice, the script formats, the hook patterns and the content pillars. When a raw thought goes in, the rules are already baked in, so what comes back out is native to each format rather than a generic blob that still needs hours of work.
The gap between idea and published has to stay short. If a thought sits in a drafts folder for more than a day it loses the energy that made it worth posting in the first place.
Five minutes in CTRL App, stream of consciousness, no editor running in the back of my head. It might be a frustration with a tool I am using, an insight from yesterday's build, a story from the cyber years or a contrarian take that has been rattling around for a week. Format is not the job at this stage. The job is just getting the thought out of my head and onto something I can work with later.
That raw capture goes straight into the becomingscott project inside Claude. The project has eleven custom skill files loaded against it covering brand voice, the script writer in both yapping and storytelling formats, the hook generator, the carousel builder, the caption writer and the rest. Claude already knows who I am, who the audience is and what the brand is allowed to sound like, so the conversation does not start from zero every morning.
From one raw capture, the ask is the full repurposing spread in a single pass. A yapping reel script under 40 seconds, 5 hook variations across different angles, a carousel outline with slide titles, a short caption and an email subject line. Because the skill files carry the format rules for each piece, everything comes back in the right shape without me having to explain the rules again.
Each output gets one read. The hooks get the most attention because that is where the work actually pays off, and anything that sounds too polished or too AI-written gets cut on the spot. Every script has to pass the spoken word test, which means reading it out loud and rewriting any sentence that sounds written rather than said. Then it is straight into filming or scheduling, with no second pass.
One run of this prompt covers about 60% of the daily content the brand needs. The other 40% comes from whiteboard instructionals and longer carousels that deserve a more intentional plan, and even those still start from the same morning journal capture, just routed through a different prompt with a different goal.
Without skill files, Claude will quietly produce content that sounds like every other AI-generated post drifting through the feed. With skill files loaded, it produces content that sounds like me, because the files carry the actual voice patterns, the hard rules for what to avoid (staccato fragment lists, generic motivational language and guru tone all get filtered out) and the specific format rules for each piece of content the brand ships.
Building the skill files took roughly one full focused day and they have saved hundreds of hours since. The investment is front-loaded and the returns compound from there. If you are already producing content regularly with AI and it still sounds generic, this is almost always the missing piece rather than the prompt itself.
The first one is perfecting before publishing. Raw energy is the whole point of the format, and spending 30 minutes polishing a script that was meant to feel unscripted defeats the purpose and burns the exact time the pipeline was built to save in the first place.
The second is batching too far ahead. Batching a week of content in one sitting sounds efficient until Wednesday, when the posts start feeling stale because they are no longer connected to what is actually being built that day. The pipeline works best when the input is fresh and the gap between capture and publishing is measured in hours rather than days.
The third is ignoring what already performed. The pipeline is supposed to include a feedback loop where high-performing reels get repurposed into carousels and the hooks that stopped the scroll get recycled into new formats. If something resonated with the audience, the smart move is to use it again from a different angle rather than constantly inventing new ideas from a cold start.
Book a call and we will work out the shape of it together. From there it is either work I take on directly, advisory to walk you through building your own version, or an introduction to the right person in the network. Every problem leaves the call with a path forward.
Book a Call →