Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add lsj4232/KR_PATENT_SKILL --skill docx-to-hlzgit clone --depth 1 https://github.com/lsj4232/KR_PATENT_SKILLWrote 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.
[](https://agentmods.dev/skills/lsj4232/kr_patent_skill/docx-to-hlz)<a href="https://agentmods.dev/skills/lsj4232/kr_patent_skill/docx-to-hlz"><img src="https://agentmods.dev/badge/skills/lsj4232/kr_patent_skill/docx-to-hlz/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.
<a href="https://agentmods.dev/skills/lsj4232/kr_patent_skill/docx-to-hlz"><img src="https://agentmods.dev/badge/skills/lsj4232/kr_patent_skill/docx-to-hlz.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00560 | $0.02827 |
| Opus 5.5 | $0.00224 | $0.01131 |
| Sonnet 5.5 | $0.00112 | $0.00565 |
| Haiku 4.5 | $0.00056 | $0.00283 |
Grade A, and why
docx-to-hlz 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 25d 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.
How it starts
The opening of the file, as written. The whole thing — 119 lines — stays where its author put it; the contents beside it link to each section on GitHub.
docx-to-hlz — 워드 명세서 → 특허청 편집기(.hlz) 변환
완성된 한국 특허 명세서 .docx를 특허청 전자출원 편집기가 여는 .hlz(ZIP: KIPO XML + 도면)로 변환한다.
사용법
python "<skill>/scripts/docx2hlz.py" "명세서.docx" # → 같은 폴더에 명세서.hlz
python "<skill>/scripts/docx2hlz.py" "명세서.docx" -o out.hlz
의존성: python-docx, Pillow(png/tiff 재인코딩·크기추정용).
변환 후 사용자가 편집기에서 [파일 > 열기]로 .hlz 선택(또는 창에 드래그&드롭). 편집기 GUI 클릭은 자동화 불가이므로 여는 것은 사용자가 한다.
입력 워드 전제
- KIPO 별지 서식 【식별항목】 태그로 섹션이 구분돼 있어야 한다(위 description의 12개 태그).
- 래퍼 태그(【발명의 설명】【발명의 내용】【요약서】)는 자식만 쓰고 자신은 버린다.
- 비표준 소제목(예 【1. 용어의 정의…】 탐색창용 챕터)은 본문 문단으로 흡수(대괄호 제거).
- 표(table)는 미지원(현재 대상 명세서에 표가 없어 미구현) — 표 있는 건은 확장 필요.
산출 포맷 핵심 (application-body DTD)
<KIPO keapsVersion="5.6" editorKind="K" pageCount wi imgApply="N" xmlns="http://www.kipo.go.kr">
<PatentCAFDOC docflag="1.0" documentID="숫자">
<description>
<invention-title>한글{ENGLISH}</invention-title>
<technical-field><p num="0001">…</p></technical-field>
<background-art>…</background-art>
<summary-of-invention><tech-problem/><tech-solution/><advantageous-effects/></summary-of-invention>
<description-of-drawings><p>도 1은…<br/>도 2는…</p></description-of-drawings>
<description-of-embodiments><p num=…>…</p></description-of-embodiments>
</description>
<claims><claim num="1"><claim-text>…단계;<br/>…을 특징으로 하는 ….</claim-text></claim></claims>
<abstract><summary><p num="0001a"/></summary><abstract-figure><figref num="3"/></abstract-figure></abstract>
<drawings><figure num="1"><img id he wi file="pat00001.tif" img-format="tif"/></figure></drawings>
</PatentCAFDOC>
</KIPO>
편집기 호환 하드 규칙 (스크립트가 자동 적용 — 하나라도 어기면 로딩 실패)
- ZIP 파일명 = CP949 + UTF-8 플래그(0x800) 제거. ★가장 잘 틀리는 부분★
Python 기본은 한글 파일명에 UTF-8 플래그를 세팅 → 구 MFC 편집기가 엔트리를 못 찾아
압축해제 자체가 실패(temp 폴더 빔) → "…SGM 을(를) 찾을 수 없습니다" 에러.
ZipInfo._encodeFilenameFlags오버라이드로 해결. 레퍼런스 hlz는 전 엔트리 flag=0x0000. - 내부 XML basename == .hlz 파일 basename (편집기가 이름으로 문서파일 탐색).
- img-format ∈ {jpg, tif} (DTD: jpg|tif|st33|st35). tiff→tif(3글자), png/bmp/gif→jpg 재인코딩(투명→흰색).
- 청구항은
<claim-text>한 덩어리 +<br/>. 【청구항 N】 뒤 작성자 메모//…//는 제거. <p num>= description 전관통 4자리 일련번호. abstract만 별도 카운터 + 접미사a.- 대표도 =
<abstract-figure><figref num>. 도면 =<drawings><figure num><img he wi file img-format>, he/wi=mm(EMU/36000, 최대 165×222mm·E-218). - XML 특수문자 이스케이프, UTF-8. ZIP은 XML 먼저.
What ships with it
2 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 25d ago First seen · 119 lines · 560 tokens per session scan A 94d1a3ca8707
docx-to-hlz is a skill published in the GitHub repository lsj4232/KR_PATENT_SKILL (22 stars, last pushed 5d ago), licensed MIT. It adds 560 tokens to every session and 2,827 once invoked, about $0.0022 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 2026-09-22.
Other skills, from other repositories
court-sms
A workflow that reads court text messages or delivery links, extracts case details, downloads court documents, and files them in the matching case folder.
recht-report-pdf-momarcode1
Generates professional PDF reports with risk scores, cost estimates, and recommendations using ReportLab. Austrian law branding.
law-to-markdown
A legal-document conversion skill for turning PDF, Word, or plain-text regulations into structured Markdown. Markdown is a text format that uses simple symbols for headings and lists.
chinese-copyright-application
A document-generation toolkit for preparing Chinese software copyright registration materials from a software project’s code and documentation. It supports applications for web apps, mobile apps, desktop software, and WeChat mini programs.
chinese-copyright-application-okooo5km
A document-generation tool for preparing Chinese software copyright registration materials from a software project. It creates a Markdown application form and LaTeX source files for a source-code document, user manual, and design description.
waliduj-podpis-pdf-pl
Sprawdza podpis elektroniczny w PRZYCHODZĄCYM PDF - czy plik zmieniono po podpisaniu, kto go podpisał, kiedy, czy jest znacznik czasu i czy podpis obejmuje cały dokument. Działa lokalnie na pyHanko, bez sieci i bez wysyłania pliku gdziekolwiek (RODO-safe, tajemnica zawodowa). Wyłącznie odczyt - NIE podpisuje niczego.…