技能潮 SkillTide 提交收录
官方说明2 个来源Agent 技能3 天前
nuncaeslupus/literary-critic

nuncaeslupus/literary-critic

Lets an LLM index, analyse and advise on a whole novel manuscript like a professional editor, across drafts
在 GitHub 查看 ↗直链 · 无中转页
★ 数取值
0 ★ · nuncaeslupus/literary-critic · 取自 GitHub 搜索 API · (2026-10-16 20:00:20 核对)
来源明细
从该仓库的 GitHub 页面直接读取 · 条目类型:开源项目
收录与活跃
收录于 2026-10-14 · 最近见到 2026-10-16 · 共见到 3 次 · 被 2 个来源列出
评分构成
综合分 203 · 2 个来源 · repo
常见问题(答案都来自我们的记录)
怎么安装?
我们【没有】从该仓库取到安装命令(我们的 install 覆盖率只有约 4%)。请看它的 README 或 releases 页。
它适合谁 / 用来做什么?
我们把它归在这些标签下: Agent 技能。标签来自条目自身,不是我们强加的分类。
谁维护它、还在更新吗?
仓库 nuncaeslupus/literary-critic;我们收录于 2026-10-14;最近见到 2026-10-16。这里没有"官方认证"或"质量评级"——我们只记录来源与时间。
第三方质量标记

literary-critic

以下摘录自该仓库的 README(片段) · 在 GitHub 读完整 README →

Tooling that lets an LLM work through a full novel manuscript as a professional
editor would — and that keeps working across many chat sessions and drafts.
1. **Index** — turn the manuscript into stable anchors (part → chapter → scene →
paragraph) and a Codex — a knowledge base of scenes, characters, places,
facts, setups/payoffs and findings, each tied to its text — so the LLM can
find and relate anything without loading the whole book.
2. **Analyse** — run a generic catalogue of checks (structure, characters, plot,
scenes, POV, prose, dialogue, continuity, …), adapted to each novel, with
evidence anchored to the text and script-measured metrics wherever possible.
3. **Advise** — produce prioritised, concrete revision actions in rounds
(structural first, then scene, then line level), track their status, re-index
the next draft and measure what changed.
For novelists, and the editors who work with them, who want an LLM's reading of
a whole manuscript without handing it the pen. In Claude Code:
Status: early development. See status/specification.md and status/plan.md.
Documents are read and written through
doc-control, which owns the generic
document work (paragraph addresses, chapters and scenes, diff, locating,
edits with a backup, comments, undo). literary-critic owns the literary side: the
index, the Codex, the checks, letters and revision rounds. See
docs/architecture.md.
How it is used
| Where | How |
|-------|-----|
| Claude Code (local or cloud) | Clone, open, talk. Scripts run directly. |
| Claude (desktop app with Cowork, or Claude.ai) | Add the literary-critic Skill once (lc kit --skill builds the zip to upload in Claude's settings, under Skills), then say "I want to use literary-critic for my novel". |
| Chat with code execution (ChatGPT, Claude.ai, Gemini) | Upload the kit zip (lc kit builds it; the assistant reads its START_HERE.md) and your workspace zip; the assistant runs the scripts and returns an updated workspace zip. |
| Chat without code execution | Paste the portable prompts from prompts/; reduced mode without metrics. |
**What you need for the full experience** (the novel's folder connected, changes
written into the file with a backup, Word opened at the right paragraph): Claude's
desktop app with Cowork, which needs a paid Claude plan and a Mac with Apple
silicon or Windows Pro, Enterprise or Education. On the free plan, Windows Home
or an Intel Mac there is no Cowork: the tool works in basic mode, where you
upload the manuscript and get back a commented copy.
Your novel and all working files live in a **workspace folder outside this
来源
github · https://github.com/nuncaeslupus/literary-critic
按领域浏览
Agent 技能
收录于
昨天

更多Agent 技能

查看Agent 技能全部 142 条 →

所属领域与标签

每个标签都能点进去看同类的全部条目。数字是当前库里的真实条数。

浏览全部领域 →

← 返回索引