Humanize KR (im-not-ai)
epoko77-ai/im-not-aiA Korean-language humanizer that detects and rewrites AI-sounding text (translationese, mechanical parallelism) while preserving facts, with deterministic gates.
Recommended
Check first Depth 4/5
Anti-Defensive Writing Claude skills
A lightweight CN/EN academic-writing skill that restructures paper drafts around one strongest claim instead of hedged, evenly-weighted self-critique.
Our verdict Recommended with caveats
A well-articulated structural editing tool for paper narrative, but its own rules (never volunteer weakness, don't phrase a non-leading result as a loss) push toward a more one-sided presentation than many venues and reviewers expect — use it for structure and framing, not as license to omit real limitations.
What Anti-Defensive Writing does
This skill gives the agent a single editorial stance for academic writing, which it calls the press-release principle: treat the paper as announcing one clear strength rather than filing an even-handed project report. Its rules tell the agent to organize the paper only around the work's genuine advantages, drop a chronological 'we tried A, then B, then C' narrative in favor of problem-gap-solution-evidence, avoid building the paper's argument on a metric where it does not win, state the advantage explicitly rather than expecting reviewers to find it in a table, limit comparison scope to claims the evidence actually supports, and restructure the whole narrative if the results don't support the original framing. A language-rules section lists self-weakening phrases to avoid ('unfortunately', 'only', 'still clearly behind') and tells the agent not to phrase a non-leading result as a loss. It covers abstracts, introductions, conclusions, cutting paper length, organizing experiments around an argumentative role, and a pre-submission checklist phrased as 'don't hand the reviewer a knife.'
The repository is tiny: two SKILL.md files (Chinese and English, no code) plus copy-paste prompt files for use with any AI assistant, and a Codex plugin wrapper with no MCP server or extra runtime dependency.
Good for
- Restructuring a paper draft, abstract or conclusion around its strongest result
- Deciding what to cut when a paper is over length
- Preparing a rebuttal without over-conceding to likely reviewer objections
Not for
- Writing that should disclose limitations fully and evenly, such as clinical, safety or policy-relevant research
- Anyone who wants the agent to flag weaknesses as prominently as strengths
What's inside depth 2/5
Two skill files (Chinese and English) averaging 6 KB with copy-paste prompt variants; no scripts, references or tests.
- Skills
- 2
- Reference docs
- 0
- Scripts
- 0
- Tests / evals
- 0
- Templates
- 0
- Avg SKILL.md
- 5.8 KB
Counted from the repository file tree. How we score depth.
Security review
No issues foundWe scanned all 14 files in this repository and read every script. Nothing in it reaches beyond what the skill describes.
Writing-instruction skills (Chinese and English) for restructuring academic papers around their strongest supported claim. The only script is a packaging helper that zips the plugin files, and PRIVACY.md states the package has no backend, telemetry or network calls, which matches the files.
What we checked
The repository (14 files) on Oct 29, 2026. We had every file and read every script.
NVIDIA SkillSpector static scan: 9 pattern hits (Agent Snooping 9). All nine hits are 'Skill Enumeration' matches on README links to the two SKILL.md files. No network code, credentials, allowed-tools or hooks.
A review, not a certification. Skills change after we check them — read the files before installing.
License
MIT — commercial use allowed.
Can it run unattended?
Not really. It is an interactive editing and restructuring aid applied to a draft, not a standing process.
Alternatives to Anti-Defensive Writing
Academic Research Skills (ARS)
Imbad0202/academic-research-skillsFive very detailed Claude Code skills for the academic pipeline: deep research, paper writing, peer review, revision and systematic-review screening.
With caveats
Check first Depth 5/5
Scientific Agent Skills
K-Dense-AI/scientific-agent-skillsK-Dense's library of 177 science skills covering bioinformatics, chemistry, clinical research, data analysis, lab tooling and scientific writing.
With caveats
Broad access Depth 4/5
FAQ
What is Anti-Defensive Writing?
It is a skill that restructures academic paper drafts around their single strongest claim, following what it calls the press-release principle, instead of an evenly hedged, self-critical structure. It comes in Chinese and English versions.
How do I install it?
Copy the skills/anti-defensive-writing (or the -en folder for English) directory into your agent's skills directory, such as ~/.claude/skills/. A copy-paste prompt version and a Codex plugin install are also provided.
Does this make a paper's claims less honest?
The skill's own rules tell the agent not to volunteer weaknesses and not to phrase a non-leading result as a loss, which goes beyond normal concise writing advice. Use it for structure and emphasis, and still disclose real limitations your venue or field expects to see.
Is it only for Chinese-language papers?
No. Both a Chinese (anti-defensive-writing) and an English (anti-defensive-writing-en) version are included, with matching rules and quick-start prompt files for each language.
Sources: GitHub repository (stars, license, file tree, last push Oct 27, 2026). Review based on the README and SKILL.md files as published.