Humanizer

Remove signs of AI-generated writing from text. Use when editing or reviewing text to make it sound more natural and human-written. Based on Wikipedia's comprehensive "Signs of AI writing" guide. Detects and fixes patterns including: inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em dash overuse, rule of three, AI vocabulary words, negative parallelisms, and excessive conjunctive phrases. The skill is an instruction-only text-editing tool whose declared requirements and instructions match its stated purpose; it requests no credentials, installs, or unusual system access—though a couple of minor inconsistencies and a broad allowed-tools list are worth noting. This skill appears internally consistent and low-risk: it only provides instructions for editing text and asks for no credentials or installs. Before installing, note the small metadata mismatch (version numbers differ) and that the skill's allowed-tools include file read/write/grep/glob — which could let an agent open or edit documents. If you plan to use it on sensitive files, require the agent to ask you before accessing or modifying files and try it first on non-sensitive sample text. Also prefer skills with a known source/homepage; consider verifying the humanizer's origin or testing outputs to ensure it preserves meaning and tone appropriately.

• • Run a minimal workflow with Humanizer and verify output quality + runtime logs.
• • Validate install command, load behavior, and rollback in a clean environment.

• • Build a focused agent workflow using Humanizer as one capability node.
• • Combine this skill with a retrieval/search skill and compare output quality.