Token导航 LogoToken导航

Agent Skill · 开发

writer-memory(作家记忆)

writer-memory 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

writer-memory 是开发类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为只读;许可证信息还不明确,第三方扫描已通过,按安装量在开发类的 21,746 个 Skill 里排第 4,943。

总安装 7,268周安装 294GitHub Stars 31,963下载 2,281核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:writer-memory(作家记忆)
来源仓库:https://github.com/yeachan-heo/oh-my-claudecode
仓库路径:skills/writer-memory
安装命令:
npx skills add https://github.com/yeachan-heo/oh-my-claudecode --skill writer-memory
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/yeachan-heo/oh-my-claudecode --skill writer-memory

这个 Skill 是做什么的

开发类 Skill 面向写代码、改代码和排查问题,一般会读写仓库文件,部分还会执行构建或测试命令。

writer-memory 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合围绕仓库状态或协作事项进行整理。适用于代码协作与项目管理场景,可结合来源仓库和原始 README 核验具体用法。

通过 npx skills add 命令从指定 GitHub 仓库安装,需确认权限范围和维护状态。安装前建议确认是否会触发联网、命令执行或文件读写,避免误操作生产数据。适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

来源仓库内路径:skills/writer-memory

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/yeachan-heo/oh-my-claudecode --skill writer-me…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

writer-memory 怎么安装?

在终端执行 npx skills add https://github.com/yeachan-heo/oh-my-claudecode --skill writer-memory。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

writer-memory 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

writer-memory 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

writer-memory 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

writer-memory 用的人多吗?

公开数据是累计安装 7,268 次,近一周安装 294 次,下载 2,281 次,GitHub 31,963 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

writer-memory 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。
  • 来源仓库有 31,963 星,维护活跃度相对可查。
  • 标注为只读权限,试用阶段不容易误改本地文件。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理开发类任务、又不想每次都手写提示词的人。
  • 只想先试试效果、不希望它改动本地文件的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

开发类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
skill-vetter

技能审查员 · OpenClaw

8,779,018986操作浏览器MIT-0通过
ontology

本体论 · OpenClaw

6,636,960576只读MIT-0通过
entra-app-registration

内部应用程序注册 · Codex、Claude

6,631,872764接触敏感数据未明确通过
entra-app-registration

内部应用程序注册 · Codex、Claude

3,757,680201接触敏感数据未明确通过
nano-banana-pro

纳米香蕉专业版 · OpenClaw

3,526,898364接触敏感数据MIT-0有提醒
writer-memory

作家记忆 · 本页

7,26831,963只读未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Writer Memory - Agentic Memory System for Writers

Persistent memory system designed for creative writers, with first-class support for Korean storytelling workflows.

Overview

Writer Memory maintains context across Claude sessions for fiction writers. It tracks:

  • Characters (캐릭터): Emotional arcs (감정궤도), attitudes (태도), dialogue tone (대사톤), speech levels
  • World (세계관): Settings, rules, atmosphere, constraints
  • Relationships (관계): Character dynamics and evolution over time
  • Scenes (장면): Cut composition (컷구성), narration tone, emotional tags
  • Themes (테마): Emotional themes (정서테마), authorial intent

All data persists in .writer-memory/memory.json for git-friendly collaboration.

Commands

CommandAction
/oh-my-claudecode:writer-memory init <project-name>Initialize new project memory
/oh-my-claudecode:writer-memory statusShow memory overview (character count, scene count, etc)
/oh-my-claudecode:writer-memory char add <name>Add new character
/oh-my-claudecode:writer-memory char <name>View character details
/oh-my-claudecode:writer-memory char update <name> <field> <value>Update character field
/oh-my-claudecode:writer-memory char listList all characters
/oh-my-claudecode:writer-memory rel add <char1> <char2> <type>Add relationship
/oh-my-claudecode:writer-memory rel <char1> <char2>View relationship
/oh-my-claudecode:writer-memory rel update <char1> <char2> <event>Add relationship event
/oh-my-claudecode:writer-memory scene add <title>Add new scene
/oh-my-claudecode:writer-memory scene <id>View scene details
/oh-my-claudecode:writer-memory scene listList all scenes
/oh-my-claudecode:writer-memory theme add <name>Add theme
/oh-my-claudecode:writer-memory world set <field> <value>Set world attribute
/oh-my-claudecode:writer-memory query <question>Query memory naturally (Korean supported)
/oh-my-claudecode:writer-memory validate <character> <dialogue>Check if dialogue matches character tone
/oh-my-claudecode:writer-memory synopsisGenerate emotion-focused synopsis
/oh-my-claudecode:writer-memory exportExport full memory as readable markdown
/oh-my-claudecode:writer-memory backupCreate manual backup

Memory Types

캐릭터 메모리 (Character Memory)

Tracks individual character attributes essential for consistent portrayal:

FieldKoreanDescription
arc감정궤도Emotional journey (e.g., "체념 -> 욕망자각 -> 선택")
attitude태도Current disposition toward life/others
tone대사톤Dialogue style (e.g., "담백", "직설적", "회피적")
speechLevel말투 레벨Formality: 반말, 존댓말, 해체, 혼합
keywords핵심 단어Characteristic words/phrases they use
taboo금기어Words/phrases they would never say
emotional_baseline감정 기준선Default emotional state
triggers트리거What provokes emotional reactions

Example:

/writer-memory char add 새랑
/writer-memory char update 새랑 arc "체념 -> 욕망자각 -> 선택"
/writer-memory char update 새랑 tone "담백, 현재충실, 감정억제"
/writer-memory char update 새랑 speechLevel "해체"
/writer-memory char update 새랑 keywords "그냥, 뭐, 괜찮아"
/writer-memory char update 새랑 taboo "사랑해, 보고싶어"

세계관 메모리 (World Memory)

Establishes the universe your story inhabits:

FieldKoreanDescription
setting배경Time, place, social context
rules규칙How the world operates (magic systems, social norms)
atmosphere분위기Overall mood and tone
constraints제약What cannot happen in this world
history역사Relevant backstory

관계 메모리 (Relationship Memory)

Captures the dynamic between characters over time:

FieldDescription
typeBase relationship: romantic, familial, friendship, rivalry, professional
statusCurrent state: budding, stable, strained, broken, healing
power_dynamicWho has the upper hand, if any
eventsTimeline of relationship-changing moments
tensionCurrent unresolved conflicts
intimacy_levelEmotional closeness (1-10)

Example:

/writer-memory rel add 새랑 해랑 romantic
/writer-memory rel update 새랑 해랑 "첫 키스 - 새랑 회피"
/writer-memory rel update 새랑 해랑 "해랑 고백 거절당함"
/writer-memory rel update 새랑 해랑 "새랑 먼저 손 잡음"

장면 메모리 (Scene Memory)

Tracks individual scenes and their emotional architecture:

FieldKoreanDescription
title제목Scene identifier
characters등장인물Who appears
location장소Where it happens
cuts컷 구성Shot-by-shot breakdown
narration_tone내레이션 톤Narrative voice style
emotional_tag감정 태그Primary emotions (e.g., "설렘+불안")
purpose목적Why this scene exists in the story
before_after전후 변화What changes for characters

테마 메모리 (Theme Memory)

Captures the deeper meaning woven through your story:

FieldKoreanDescription
name이름Theme identifier
expression표현 방식How this theme manifests
scenes관련 장면Scenes that embody this theme
character_links캐릭터 연결Which characters carry this theme
author_intent작가 의도What you want readers to feel

Synopsis Generation (시놉시스)

The /synopsis command generates an emotion-focused summary using 5 essential elements:

5 Essential Elements (시놉시스 5요소)

  1. 주인공 태도 요약 (Protagonist Attitude Summary)

- How the protagonist approaches life/love/conflict - Their core emotional stance - Example: "새랑은 상실을 예방하기 위해 먼저 포기하는 사람"

  1. 관계 핵심 구도 (Core Relationship Structure)

- The central dynamic driving the story - Power imbalances and tensions - Example: "사랑받는 자와 사랑하는 자의 불균형"

  1. 정서적 테마 (Emotional Theme)

- The feeling the story evokes - Not plot, but emotional truth - Example: "손에 쥔 행복을 믿지 못하는 불안"

  1. 장르 vs 실제감정 대비 (Genre vs Real Emotion Contrast)

- Surface genre expectations vs. actual emotional content - Example: "로맨스지만 본질은 자기수용 서사"

  1. 엔딩 정서 잔상 (Ending Emotional Aftertaste)

- The lingering feeling after the story ends - Example: "씁쓸한 안도, 불완전한 해피엔딩의 여운"

Character Validation (캐릭터 검증)

The /validate command checks if dialogue matches a character's established voice.

What Gets Checked

CheckDescription
Speech LevelDoes formality match? (반말/존댓말/해체)
Tone MatchDoes the emotional register fit?
Keyword UsageUses characteristic words?
Taboo ViolationUses forbidden words?
Emotional RangeWithin character's baseline?
Context FitAppropriate for relationship and scene?

Validation Results

  • PASS: Dialogue is consistent with character
  • WARN: Minor inconsistencies, may be intentional
  • FAIL: Significant deviation from established voice

Example:

/writer-memory validate 새랑 "사랑해, 해랑아. 너무 보고싶었어."

Output:

[FAIL] 새랑 validation failed:
- TABOO: "사랑해" - character avoids direct declarations
- TABOO: "보고싶었어" - character suppresses longing expressions
- TONE: Too emotionally direct for 새랑's 담백 style

Suggested alternatives:
- "...왔네." (minimal acknowledgment)
- "늦었다." (deflection to external fact)
- "밥 먹었어?" (care expressed through practical concern)

Context Query (맥락 질의)

Natural language queries against memory, with full Korean support.

Example Queries

/writer-memory query "새랑은 이 상황에서 뭐라고 할까?"
/writer-memory query "규리의 현재 감정 상태는?"
/writer-memory query "해랑과 새랑의 관계는 어디까지 왔나?"
/writer-memory query "이 장면의 정서적 분위기는?"
/writer-memory query "새랑이 먼저 연락하는 게 맞아?"
/writer-memory query "해랑이 화났을 때 말투는?"

The system synthesizes answers from all relevant memory types.

Behavior

  1. On Init: Creates .writer-memory/memory.json with project metadata and empty collections
  2. Auto-Backup: Changes are backed up before modification to .writer-memory/backups/
  3. Korean-First: Emotion vocabulary uses Korean terms throughout
  4. Session Loading: Memory is loaded on session start for immediate context
  5. Git-Friendly: JSON formatted for clean diffs and collaboration

Integration

With OMC Notepad System

Writer Memory integrates with .omc/notepad.md:

  • Scene ideas can be captured as notes
  • Character insights from analysis sessions are preserved
  • Cross-reference between notepad and memory

With Architect Agent

For complex character analysis:

Task(subagent_type="oh-my-claudecode:architect",
     model="opus",
     prompt="Analyze 새랑's arc across all scenes...")

Character Validation Pipeline

Validation pulls context from:

  • Character memory (tone, keywords, taboo)
  • Relationship memory (dynamics with dialogue partner)
  • Scene memory (current emotional context)
  • Theme memory (authorial intent)

Synopsis Builder

Synopsis generation aggregates:

  • All character arcs
  • Key relationship events
  • Scene emotional tags
  • Theme expressions

Examples

Full Workflow

# Initialize project
/writer-memory init 봄의 끝자락

# Add characters
/writer-memory char add 새랑
/writer-memory char update 새랑 arc "체념 -> 욕망자각 -> 선택"
/writer-memory char update 새랑 tone "담백, 현재충실"
/writer-memory char update 새랑 speechLevel "해체"

/writer-memory char add 해랑
/writer-memory char update 해랑 arc "확신 -> 동요 -> 기다림"
/writer-memory char update 해랑 tone "직진, 솔직"
/writer-memory char update 해랑 speechLevel "반말"

# Establish relationship
/writer-memory rel add 새랑 해랑 romantic
/writer-memory rel update 새랑 해랑 "첫 만남 - 해랑 일방적 호감"
/writer-memory rel update 새랑 해랑 "새랑 거절"
/writer-memory rel update 새랑 해랑 "재회 - 새랑 내적 동요"

# Set world
/writer-memory world set setting "서울, 현대, 20대 후반 직장인"
/writer-memory world set atmosphere "도시의 건조함 속 미묘한 온기"

# Add themes
/writer-memory theme add "포기하지 않는 사랑"
/writer-memory theme add "자기 보호의 벽"

# Add scene
/writer-memory scene add "옥상 재회"

# Query for writing
/writer-memory query "새랑은 이별 장면에서 어떤 톤으로 말할까?"

# Validate dialogue
/writer-memory validate 새랑 "해랑아, 그만하자."

# Generate synopsis
/writer-memory synopsis

# Export for reference
/writer-memory export

Quick Character Check

/writer-memory char 새랑

Output:

## 새랑

**Arc (감정궤도):** 체념 -> 욕망자각 -> 선택
**Attitude (태도):** 방어적, 현실주의
**Tone (대사톤):** 담백, 현재충실
**Speech Level (말투):** 해체
**Keywords (핵심어):** 그냥, 뭐, 괜찮아
**Taboo (금기어):** 사랑해, 보고싶어

**Relationships:**
- 해랑: romantic (intimacy: 6/10, status: healing)

**Scenes Appeared:** 옥상 재회, 카페 대화, 마지막 선택

Storage Schema

{
  "version": "1.0",
  "project": {
    "name": "봄의 끝자락",
    "genre": "로맨스",
    "created": "2024-01-15T09:00:00Z",
    "lastModified": "2024-01-20T14:30:00Z"
  },
  "characters": {
    "새랑": {
      "arc": "체념 -> 욕망자각 -> 선택",
      "attitude": "방어적, 현실주의",
      "tone": "담백, 현재충실",
      "speechLevel": "해체",
      "keywords": ["그냥", "뭐", "괜찮아"],
      "taboo": ["사랑해", "보고싶어"],
      "emotional_baseline": "평온한 무관심",
      "triggers": ["과거 언급", "미래 약속"]
    }
  },
  "world": {
    "setting": "서울, 현대, 20대 후반 직장인",
    "rules": [],
    "atmosphere": "도시의 건조함 속 미묘한 온기",
    "constraints": [],
    "history": ""
  },
  "relationships": [
    {
      "id": "rel_001",
      "from": "새랑",
      "to": "해랑",
      "type": "romantic",
      "dynamic": "해랑 주도 → 균형",
      "speechLevel": "반말",
      "evolution": [
        { "timestamp": "...", "change": "첫 만남 - 해랑 일방적 호감", "catalyst": "우연한 만남" },
        { "timestamp": "...", "change": "새랑 거절", "catalyst": "과거 트라우마" },
        { "timestamp": "...", "change": "재회 - 새랑 내적 동요", "catalyst": "옥상에서 재회" }
      ],
      "notes": "새랑의 불신 vs 해랑의 기다림",
      "created": "..."
    }
  ],
  "scenes": [
    {
      "id": "scene-001",
      "title": "옥상 재회",
      "characters": ["새랑", "해랑"],
      "location": "회사 옥상",
      "cuts": ["해랑 먼저 발견", "새랑 굳은 표정", "침묵", "해랑 먼저 말 걸기"],
      "narration_tone": "건조체",
      "emotional_tag": "긴장+그리움",
      "purpose": "재회의 어색함과 남은 감정 암시",
      "before_after": "새랑: 무관심 -> 동요"
    }
  ],
  "themes": [
    {
      "name": "포기하지 않는 사랑",
      "expression": "해랑의 일관된 태도",
      "scenes": ["옥상 재회", "마지막 고백"],
      "character_links": ["해랑"],
      "author_intent": "집착이 아닌 믿음의 사랑"
    }
  ],
  "synopsis": {
    "protagonist_attitude": "새랑은 상실을 예방하기 위해 먼저 포기하는 사람",
    "relationship_structure": "기다리는 자와 도망치는 자의 줄다리기",
    "emotional_theme": "사랑받을 자격에 대한 의심",
    "genre_contrast": "로맨스지만 본질은 자기수용 서사",
    "ending_aftertaste": "불완전하지만 따뜻한 선택의 여운"
  }
}

File Structure

.writer-memory/
├── memory.json          # Main memory file
├── backups/             # Auto-backups before changes
│   ├── memory-2024-01-15-090000.json
│   └── memory-2024-01-20-143000.json
└── exports/             # Markdown exports
    └── export-2024-01-20.md

Tips for Writers

  1. Start with Characters: Build character memories before scenes
  2. Update Relationships After Key Scenes: Track evolution actively
  3. Use Validation While Writing: Catch voice inconsistencies early
  4. Query Before Difficult Scenes: Let the system remind you of context
  5. Regular Synopsis: Generate periodically to check thematic coherence
  6. Backup Before Major Changes: Use /backup before significant story pivots

Troubleshooting

Memory not loading?

  • Check .writer-memory/memory.json exists
  • Verify JSON syntax is valid
  • Run /writer-memory status to diagnose

Validation too strict?

  • Review taboo list for unintended entries
  • Consider if character is growing (arc progression)
  • Intentional breaks from pattern are valid for dramatic moments

Query not finding context?

  • Ensure relevant data is in memory
  • Try more specific queries
  • Check character names match exactly

平台分布

Codex

32.94%
按下载量换算751

Claude

30.98%
按下载量换算707

Cursor

17.66%
按下载量换算403

Gemini CLI

9.45%
按下载量换算216

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills