Claude Skill MartBrowse skillsQuick linesLearn by videoTerminal guideWhat is a Skill?
 
← Back to list

Academic Humanizer

Minimally rewrites or audits English/Chinese academic prose to remove formulaic, vacuous, and process-leaking language without altering claims, numbers, or terminology.

Content & WritingIntermediate★ 738⑂ 60AI score 9/10Last updated: Aug 26, 2026

What it does

Targets observable writing defects in scholarly prose — empty evaluation, mechanical repetition, and process leakage — instead of imitating human imperfection or gaming a detector.

  • Four operations: generate, detect, rewrite, edit
  • A single C0–C7 arbitration table: artifact boundary > semantic fidelity > locked spans (quotes, formulas, code, citations) > terminology identity > academic register > argument structure > document patterning > local style repair
  • Internal claim/evidence ledger and terminology ledger prevent adding facts, citations, baselines, or limitations that the source never contained
  • Builds a document-wide distribution map (dashes, parallel triads, paragraph templates, sentence rhythm) and re-verifies it after editing
  • Explicit stop conditions: missing facts, unverifiable citations, excerpt-vs-manuscript mismatch, or requests for detector evasion

Who it's for

  • Graduate students and researchers polishing papers, abstracts, grants, and reviewer responses
  • Authors who want LLM-draft phrasing cleaned up with zero meaning drift
  • Editors maintaining consistent terminology across bilingual (EN/ZH) manuscripts

Examples

  1. "De-AI this Discussion paragraph" → strips unsupported certainty and hollow evaluative sentences while keeping every number, hedge, and causal claim; returns only the clean text.
  2. "Don't rewrite — just flag AI-sounding phrasing" → detect mode returns findings with exact quotes, rule IDs, distribution evidence, and a change/keep/uncertain verdict.
  3. A coined method name drifts across abstract, body, and figure captions → the terminology gate normalizes it to the canonical term while preserving declared abbreviations and intentional distinctions.

· · · Install guide · · ·

Try it now, no install

Paste this into Claude to use the skill without installing anything.

Read the instructions in this file and follow them to help me:
https://raw.githubusercontent.com/dongshuyan/compass-skills/HEAD/skills/academic-humanizer/SKILL.md

What I want: (describe your task here)

If Claude can't open the link, open it yourself and paste the contents instead.

↓ If it works for you, download the ZIP below and install it. Then it runs on its own — no pasting each time.

Install in the Claude app (no terminal)
  1. Download the ZIP with the button below.
  2. In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
  3. Go to Customize → Skills → + → 'Upload a skill' and upload the ZIP.
↓ Download ZIP
Install in Claude Code

Let Claude do it — paste this into Claude Code

Install the skill I found on Claude Skill Mart.
Copy the skills/academic-humanizer folder from the GitHub repo dongshuyan/compass-skills into my ~/.claude/skills/academic-humanizer/.
When it's done, tell me in one line what this skill can do.

Install with a command instead

git clone https://github.com/dongshuyan/compass-skills.git /tmp/compass-skills && mkdir -p ~/.claude/skills && cp -r /tmp/compass-skills/skills/academic-humanizer ~/.claude/skills/

⚠ This is a third-party skill. Check the source repository before installing.

  1. Open a terminal.
  2. Clone the repo: git clone https://github.com/dongshuyan/compass-skills.git
  3. Create the skills directory: mkdir -p ~/.claude/skills
  4. Copy the skill: cp -r compass-skills/skills/academic-humanizer ~/.claude/skills/
  5. Confirm that references/ and scripts/metrics.py came along — the core rules live in those files.
  6. Restart Claude Code and try a prompt like "humanize this abstract" or "audit this section for AI-like phrasing".
  7. Optional: install Python 3 if you want the reproducible metrics script; the skill works fine without it.
View source on GitHub ↗License: MIT