All products

Learn text revision with AI

See each edit alongside its defect, concrete test and effect on the reader

Learn text revision with AI by examining why an edit works. The teach-revision skill selects a few valuable teaching points from your own draft. Each move identifies a defect, shows the concrete test that exposed it and explains the reader effect behind the proposed fix.

Before-and-after wording makes the process visible so you can try the test elsewhere. Claims, numbers, conditions and your voice must survive the revision. The model’s checklist and local indicators support review; they do not certify semantic equivalence, objectively grade the writing or prove that you have mastered the method.

What it does

  • Inspect staged revisions of your own draft
  • Connect each change to a reader effect
  • Compare meaning before and after an edit
  • Practise reusable checks on another sentence
  • Distinguish structural and sentence-level issues
  • Focus on a manageable set of teaching points

Install and use

  1. Prepare the complete supported package, including methodology and tools, in a compatible environment such as Claude Code or Codex. A lone skill folder is insufficient; local helpers need Node.js 18 or newer.
  2. Provide your draft, level and intended reader. For example: “Teach me to revise this report paragraph, showing an ambiguity test and its fix while preserving every figure.”
  3. Compare each move with the original and try its test on another sentence. Ask for a simpler explanation when needed rather than teaching every defect at once.
  4. Review the improved draft and change record. Keep meaning changes or unresolved questions open for your decision.

FAQ

Source

Made by OC-NeuralSense. Original skill: teach-revision, published on GitHub under the Apache-2.0 licence.

View on GitHub