Repurposing Across Platforms with an AI Content Skill Graph

Definition

A content production method that splits brand voice, customers, per-channel rules, hooks, repurposing order and schedule into small markdown files, links them with wikilinks, and then has an AI agent read the relevant files before it works. Instead of taking one shared draft and only changing its size, you re-decide the angle, hook, tone, format and depth for every platform.

Perspectives

Ronin (2026-04-10, X)

Keep index.md as a command center that holds identity, a node map and an execution order rather than a plain file list, and start with 17 files under platforms/, voice/, engine/ and audience/. Each node links to other nodes, and before writing the agent reads the brand voice, hooks, repurposing rules and that platform's rules.

Repurposing starts on X by nailing the core idea and a short hook first, then expands in order to LinkedIn's narrative, Instagram's carousel, TikTok's short video, YouTube's long tutorial, and then the newsletter, Threads and Facebook — but rather than reformatting the same sentences, pick a different entry point for each channel.

For tools you can choose to fit how you work: Obsidian to view and edit the graph, Claude Projects to load the files as persistent context, ChatGPT to attach the key files, or Cursor to read and update the local files directly.

How to apply

• Fits a builder who already publishes on three or more platforms and keeps shipping near-duplicates — the graph's setup cost only pays back when the same idea goes out in several formats every week.
• If identical short posts are acceptable on every destination, bot-based mirroring from X is a lower-setup alternative; use this graph when each channel needs a different hook, format or depth.
• Assumes the topic inputs already exist; the graph is the output stage. Gather customer language and competitor gaps first with Collecting Content Demand with an Audience Radar, and use a product-development story from Founder-Led SaaS Pre-Demand and Launch as raw material.
• Not a starting point for a new account on one channel — the daily contribute-and-interact habits in Operating a Founder-Led Social Account come first; a single-channel account gains nothing from a repurposing order.
• Tool choice (Obsidian, Claude, ChatGPT, Cursor) follows how you already work; none is required. Channel context for the whole chain is in Social & Community.

Limits

• Running 10 accounts and cutting cost is the author's self-report; per-account quality, revenue contribution and human review time are not disclosed.

• The original's per-platform length, posting frequency, timing and algorithm advice are the author's recommendations that shift with the moment and the account. Don't treat them as fixed facts — test them on each channel.

• The 17 files are a starting template, not a validation of the optimal structure. As connections grow, you have to manage which node is read when and the context cost.

• Automated generation does not solve fact-checking, duplication, copyright, platform policy or scheduled-post failures. Review before publishing and source tracing are needed separately.

• Original link