im-not-ai: Instructions file for Claude Code

CLAUDE.md

im-not-ai CLAUDE.md is an instructions file for Claude Code from epoko77-ai/im-not-ai. It costs 5,360 tokens per session, scanned A, original, MIT.

A harness for rewriting Korean text produced by AI so it reads more like human writing while keeping its facts, claims, numbers, names, and quotations unchanged. It detects common AI-writing patterns and chooses a light, standard, or heavy editing route.

In plain words
What is it for?
Use it to revise Korean articles, reports, blog posts, and other prose, including long drafts that need diagnosis, targeted editing, and a final check.
Why use it?
It removes mechanical wording, translation-like phrasing, repetitive sentence patterns, and excessive formatting without changing the text’s meaning or genre.

Instructions file for Claude Code

Written for Claude Code: ${CLAUDE_SKILL_DIR} variable. Also seen: mentions .claude-plugin; mentions CLAUDE.md; mentions Codex.

This is epoko77-ai/im-not-ai's own configuration. It tells Claude Code how to work on im-not-ai itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything im-not-ai configures →

Needs its repository: it reads a path above its own folder, which exists only inside the repository. The line is │ ├── references → ../../../skills/humanize-korean/references.

About the project

Humanize KR is a command-line coding-agent skill that detects patterns making Korean text look machine-written and rewrites its style, rhythm, and wording while preserving the content. Korean-language writers use it to revise translation-like phrasing, repetitive structures, formulaic expressions, and other listed AI writing patterns. The catalogue includes its agents, skills, instructions, and plugin for supported coding-agent tools.

epoko77-ai/im-not-ai · 6,252 stars · on GitHub · imnotai.kr

Reuse

Borrowing it

Nothing to install: this file belongs to epoko77-ai/im-not-ai. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/epoko77-ai/im-not-ai/main/CLAUDE.md
Clone the repo
git clone --depth 1 https://github.com/epoko77-ai/im-not-ai

Made for: Claude Code.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for im-not-ai CLAUDE.md

README.md
[![agentmods](https://agentmods.dev/badge/instructions/epoko77-ai/im-not-ai/claude-md/github.svg)](https://agentmods.dev/instructions/epoko77-ai/im-not-ai/claude-md)
Your own site
<a href="https://agentmods.dev/instructions/epoko77-ai/im-not-ai/claude-md"><img src="https://agentmods.dev/badge/instructions/epoko77-ai/im-not-ai/claude-md/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for im-not-ai CLAUDE.md

Your own site · 80×15
<a href="https://agentmods.dev/instructions/epoko77-ai/im-not-ai/claude-md"><img src="https://agentmods.dev/badge/instructions/epoko77-ai/im-not-ai/claude-md.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 5,360 This file is loaded in full into every session.
When invoked 5,360 The same file — it is already loaded in full.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.05360 $0.05360
Opus 5.5 $0.02144 $0.02144
Sonnet 5.5 $0.01072 $0.01072
Haiku 4.5 $0.00536 $0.00536

Measured 29d ago against content hash 6162a53140d8, method: parsed. Prices are Anthropic first-party input rates as of 2026-10-07, from the pricing page.

Security

Grade A, and why

im-not-ai CLAUDE.md scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 29d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

CLAUDE.md · 214 lines

How it starts

The opening of the file, as written. The whole thing — 214 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Humanize KR — AI 한글 티 제거 하네스 (v2.3.2)

프로젝트 개요

AI(ChatGPT·Claude·Gemini 등)가 쓴 한글 텍스트를 "사람이 쓴 글처럼" 윤문해주는 3경로 하네스. 번역투·영어 인용 과다·기계적 병렬·관용구·피동태 남용·접속사 남발·리듬 균일성·이모지/불릿 과다 등 10대 카테고리 85개 AI 티 패턴(A-17 hold 1건 포함)을 탐지·분류해 내용은 한 글자도 건드리지 않고 문체·리듬·표현만 재작성한다.

v2.2부터 shim이 정량 점수로 산출하는 route_hint(light | standard | heavy) 가 디폴트 경로를 정한다(사용자 명시가 오버라이드). 글의 상태가 콜 수를 정하는 구조로, 구 "fast 1콜 / 정밀 3콜" 이분법을 대체한다.

v2.3부터 verify_gates.py가 문자율·진단 목표달성·대구 전멸·golden/수치의 4축을 검증한다. Codex도 같은 route_hint와 결정적 게이트를 사용해 light·standard·heavy 전체 경로를 제공한다.

  • light (1콜) — 잘 쓴 글(어휘 티 거의 0). 진단·finalize 생략, 보수 강도 단일 윤문. 손댈 게 거의 없으면 "이미 좋습니다" 조기 종료.
  • standard (2콜) — 보통의 AI 초안. 진단 1콜 + 겨냥 윤문 1콜. finalize 생략이 기본(결정적 변경률 게이트로 대체). 1만자급도 청킹 없이 단일 윤문 콜.
  • heavy (3+콜) — 중증 AI 슬롭 밀집 or 초장문(15,000자 초과) or 검증 증적 필요. 진단 → 윤문(shim이 청크를 2개 이상 만든 경우에만 청크 병렬) → finalize. --strict·"정밀 모드"는 이 경로를 강제.

콜 수 절감 원리 — 절감은 모델 티어가 아니라 콜 수에서 온다. 실측: 1만자 글 청킹 7콜 = 610K 토큰 vs 단일 콜 = 134K(품질 동등). 청크마다 룰북·진단이 재로드되는 비용이 폭발 원인. 따라서 단일 콜 우선, 청킹은 heavy 전용·15,000자 이하 비권장.

철칙

  1. 의미 불변 (Fidelity First) — 사실·주장·수치·고유명사·인용은 100% 원문 보존.
  2. 근거 기반 (Span-Grounded) — 모든 변경은 탐지 finding에 연결. 탐지 없는 구간은 건드리지 않음.
  3. 장르 유지 (Tone Match) — 칼럼을 문학으로, 리포트를 에세이로 옮기지 않음.
  4. 과윤문 금지 (No Over-Polish) — 변경률 30% 초과 시 경고, 50% 초과 시 강제 중단.
  5. register 보존 — 양방향 — 격식체 입력은 격식체 출력, 구어 입력은 구어 출력. 원문보다 딱딱하게 만들지 않는다: '-했-' → '-하였-' 치환 금지, '~인데요/~거든요/~한 겁니다' 구어 종결 보존. (하향 금지만 있던 기존 단방향 규칙으로는 합쇼체가 유지되는 '했→하였' 상향을 못 잡았다.) AI 티는 문법·수사이지 격식 자체가 아니다.
  6. AI 티는 빼기만 하고 넣지 않는다 (No New Tells) — 원문에 없던 상투구("기록적인 성과를 거두었다"·"괄목할 만한"·"~로 평가된다") 신규 삽입 금지. 살아있는 구어는 사람 글의 증거이므로 보존한다. 철칙 #2가 "탐지 없는 구간은 손대지 않는다"라면, #6은 "손대는 구간에도 새 AI 티를 심지 않는다" — 모순이 아니라 보완이다.

AI 워터마킹에 대한 입장 (2026-08-17 확정)

배경. Anthropic 은 2026-08-11 텍스트 워터마킹 도입을 발표했다. 방식은 Google DeepMind 의 SynthID-Text — 단어를 바꿔 심는 게 아니라, 동등하게 좋은 후보들 사이에서 뽑는 난수의 출처를 키 기반 결정값으로 바꾼다. EU AI Act 투명성 의무 대응이며 적용은 전 세계 동일. 2026-08-02 이후 출시 모델부터 적용되고, 그 이전 모델(Sonnet 5 = 6/30, Opus 5 = 7/24)에는 수개월에 걸쳐 소급 적용 중이다. 탐지 API 는 아직 공개되지 않았다.

Read the full file on GitHub · 214 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 29d ago Changed · +8 lines · +246 tokens per session 6162a53140d8
  2. 1mo ago Changed · +44 tokens per session ed497eaf636f
  3. 1mo ago First seen · 206 lines · 5,070 tokens per session scan A b4cbe0dc5f10

Subscribe to this mod's changes

im-not-ai CLAUDE.md is an instructions file published in the GitHub repository epoko77-ai/im-not-ai (6,252 stars, last pushed 14d ago), licensed MIT. It adds 5,360 tokens to every session, about $0.0214 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other instructions, from other repositories

vscode buildNext.instructions.md

Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).

microsoft/vscode · 9,715 tokens

next.js AGENTS.md

AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.

vercel/next.js · 7,581 tokens

vscode oss-third-party-notices.instructions.md

Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, cache eligibility and rollout and applying the notice (cutover).

microsoft/vscode · 5,262 tokens

langchain AGENTS.md

AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.

langchain-ai/langchain · 4,680 tokens

vscode copilot-instructions.md

Copilot instructions for microsoft/vscode, covering vs code copilot instructions, project overview, root folders, core architecture (src/ folder) and built-in extensions (extensions/ folder).

microsoft/vscode · 2,731 tokens

deno copilot-instructions.md

Copilot instructions for denoland/deno, covering deno development guide for github copilot, network access, high level overview, key directories and key files to understand first.

denoland/deno · 3,077 tokens