humanize_skill

This skill helps you rewrite copy to remove AI patterns and sound more natural while preserving meaning and intent.
  • Shell

49

GitHub Stars

1

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill richtabor/agent-skills --skill humanize

  • SKILL.md6.0 KB

Overview

This skill reviews and edits copy to remove AI-generated patterns and make text sound natural. It preserves meaning, factual claims, and coverage while eliminating mechanical phrasing, filler, and common AI vocabulary. The output includes a before/after human-likeness score, the rewritten text, and a short change summary.

How this skill works

The skill scans input for 30 defined rules across six categories: content, language, style, communication, filler, and voice. It scores the original text, rewrites problem areas to preserve intent and coverage, then rescans the rewrite to produce an after score and notes. The final deliverable includes the score report, the edited text, and a brief summary of changes.

When to use it

  • Editing AI-generated drafts
  • Polishing copy that sounds robotic or overly formal
  • Preparing technical content for a general audience
  • Cleaning marketing or product text that uses buzzwords
  • Reviewing text flagged with phrases like "sounds too AI" or "this feels machine-generated"

Best practices

  • Preserve factual claims, lists, and qualifiers; do not invent facts
  • Keep genre and audience intact; do not change tone unless requested
  • Avoid em dashes and decorative elements; use straight quotes
  • Prefer contractions and natural cadence when the genre allows
  • Run the self-check to ensure edits do not reintroduce AI patterns

Example use cases

  • Humanize a product description that uses promotional clichés
  • Edit a research summary to remove "experts say" and vague attribution
  • Polish a user guide so instructions read naturally and directly
  • Convert corporate-sounding internal comms to clear, human language
  • Refine a blog draft that overuses catalog verbs like "leverage" and "showcase"

FAQ

No. It never adds new facts, dates, quotes, or citations not present in the input or provided by the user.

Can it change tone or audience?

Only if you request a different tone. By default it preserves the original genre and audience.

How is the score calculated?

Start at 100, subtract points per violation: -3 for high-impact categories, -1 for medium. The report shows category breakdowns before and after edits.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational