Academic Humanizer
Audits and minimally revises English or Chinese academic prose to remove formulaic, hollow, and process-leaking language without altering any claim.
What it does
A disciplined editor for scholarly text — papers, abstracts, grants, cover letters, reviewer responses — that removes observable writing defects instead of imitating human imperfection or gaming a detector.
- Strips process leakage: editor narration, "in this section we will…" scaffolding, tool residue
- Triages vacuous phrasing and false contrasts: flags "not merely X but Y" constructions that carry no supported claim
- Enforces terminology identity: one canonical term per concept across body, captions, and tables, with declared abbreviations preserved
- Builds a document-level distribution map: repeated templates, dash overuse, parallel triads, flat enumeration, monotone sentence rhythm
- Runs a semantic fidelity gate: numbers, negation, causal strength, baselines, modality, scope, and locked spans stay byte-identical; no invented facts, citations, or limitations
Four operations — generate, detect, rewrite, edit — with automatic language routing based on a CJK/Latin token ratio, plus an optional Python metrics script.
Who it's for
- Graduate students and researchers polishing a manuscript before submission
- Corresponding authors cleaning up AI-drafted co-author text
- Academic editors and proofreaders who need traceable findings
- Bilingual English–Chinese writers
Examples
- Abstract cleanup: "Humanize this abstract without changing meaning" returns a clean artifact with claims, numbers, and hedging intact and clichés removed.
- Audit only: detect mode returns findings grouped by severity with exact quotes, rule IDs, distribution evidence, and a change/keep/uncertain verdict — no rewriting.
- Terminology gate: when a coined method name drifts between the body and a table caption, it normalizes to the canonical term, or stops and asks if the two labels may name different concepts.
· · · 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)
- Download the ZIP with the button below.
- In Claude, open Settings → Capabilities and turn on 'Code execution and file creation'. (one time)
- Go to Customize → Skills → + → 'Upload a skill' and upload the 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 && mkdir -p ~/.claude/skills && cp -r compass-skills/skills/academic-humanizer ~/.claude/skills/⚠ This is a third-party skill. Check the source repository before installing.
- Open a terminal.
- Clone the repository:
git clone https://github.com/dongshuyan/compass-skills.git - Create the skills directory if needed:
mkdir -p ~/.claude/skills - Copy the whole skill folder, including its
references/directory:cp -r compass-skills/skills/academic-humanizer ~/.claude/skills/ - Restart Claude Code and confirm
academic-humanizerappears in your skill list. - Try it: paste a paragraph and ask "audit this for AI-like phrasing and revise minimally without changing any claim."
- Optional: install Python 3 to run
scripts/metrics.pyfor descriptive distribution reports; without Python the rule-based contracts still apply.