agentleFS
idoforgodidoforgod/cys-bookwriting-skills/skills/voice-gate/SKILL.md

voice-gate

Skill4 starsChanged 3 months agoMITView on GitHub

저자 voice·통찰·문장 패턴 cross-cutting 가드. 모든 단계 마스터 출력에 자동 발동. M1 VoiceScorer 결정론 호출 — LLM 추론 영구 금지. 임계 0.7+ 미달 시 1회 재작성·실패 시 저자 묶음. 저자 진북 #4 (작가성·정체성) backbone.

2 other projects use a version of this.

Adopted by

Most recently taken up by

The file

Files
SKILL.md

voice-gate (cross-cutting 메타 가드)

TLDR

저자 voice 6 가중치 결정론 검증 — lib/voice_scoring.VoiceScorer.compute_voice_score() 호출. 임계 0.7+ 미달 시 1회 재작성 시도 → 실패 시 저자 묶음 escalate.

Triggers

  • 모든 단계 마스터 출력 직후 자동 발동 (cross-cutting)
  • /voice-gate verify "<text>" — 단일 텍스트 voice_score 즉시 계산
  • /voice-gate audit <chapter> — 챕터 전체 voice 감사

Detailed Methodology

1. 결정론 호출 (LLM 추론 영구 금지)

import paths
from voice_scoring import VoiceScorer
import yaml

with open(paths.book_config_path()) as f:
    config = yaml.safe_load(f.read())

scorer = VoiceScorer(voice_guidelines=config["voice_guidelines"])
breakdown = scorer.compute_voice_score(output_text)

if breakdown.weighted_total < 0.7:
    # 재작성 1회 시도 (signature 적용·preferred 치환)
    ...

2. 6 가중치 (합 = 1.00, pytest 검증 완료)

가중치값측정
vocabulary_match0.27preferred·avoid substring
sentence_rhythm0.18평균 길이·max 위반
structure_pattern0.18명제→근거→사례→통찰
signature_presence0.145 시그니처 정규식
prohibition_avoidance0.13PROH-001~005
meaning_score0.10저자 사상적 척추 4개

3. 저자 5 시그니처 패턴 (Round 4·6 수정 정규식)

  1. (.+?)가 아니라\s*(.+?)(다|이다|입니다) — "AGI는 도구가 아니라 메타도구다"
  2. 이 책은\s*(.+?)\s*아니라\s*(.+?)\s*(다룬다|쓰는) — 도입부
  3. 셋 중 하나(를|만)?\s*(선택|아니) — 결론
  4. (.+?)와\s*(.+?)\s*사이의\s*거리 — 시간·공간 좌표
  5. (두|세|N|\d+|다섯)\s*(창업자|발언자|사람).*정확히\s*같은\s*\d+\s*(배|년) — 수렴

4. 위반 시 처리

  • voice_score < 0.7 → 재작성 1회 → 실패 시 저자 묶음
  • ⑩ line-edit-master 전용: strict 임계 0.85+

저자 진북 정합

진북검증
#4 작가성·정체성✅ M1 결정론 + 5 시그니처 정규식
#7 결정론 환원✅ pytest 통과 (signature·preferred·weights sum 1.0)

출처

SPEC: specs/voice-gate.md (17 항목 풀 + 부록 G·H) 실행 모듈: lib/voice_scoring.py 설정: book-config.yaml::voice_guidelines

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.