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 agentmods add skills/antonio0720/writing-intelligence/craftnpx skills add antonio0720/writing-intelligence --skill craftgit clone --depth 1 https://github.com/antonio0720/writing-intelligenceWrote 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/antonio0720/writing-intelligence/craft)<a href="https://agentmods.dev/skills/antonio0720/writing-intelligence/craft"><img src="https://agentmods.dev/badge/skills/antonio0720/writing-intelligence/craft.svg" alt="Measured on agentmods" 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 | $0.00185 | $0.01214 |
| Opus 5 | $0.00093 | $0.00607 |
| Sonnet 5 | $0.00037 | $0.00243 |
| Haiku 4.5 | $0.00018 | $0.00121 |
Grade A, and why
writing-intelligence-craft 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 4d 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 — 75 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Writing Intelligence — The Craft Library
Author: Antonio T. Smith Jr. — Founder & CEO, Density6 LLC
License: MIT
Version: 6.0.0
Companion to: writing-intelligence — the Sovereign Meaning Runtime
What this is, and why it ships separately
This bundle is the craft corpus of Writing Intelligence — the reference material the eleven passes consult while the writing is actually being built. Genre conventions. Measured voiceprints. The anti-pattern detectors that catch prose which pattern-matches to machine writing. The positive-pattern corpus that shows what good looks like instead of only naming what bad looks like. The diagnostic scan library. The compiler references for every pass.
It ships as its own installable bundle for one reason, and the reason is arithmetic rather than taste: a skill bundle may contain at most 200 files. A bundle that exceeds the ceiling does not degrade and does not warn — it simply does not load, which would turn every install instruction in the repository into a false claim. The runtime and the library together are 229 files. Split, they are 150 and 79, and both have room to grow.
Install both. They are two halves of one system.
How to use it
The governing skill is writing-intelligence. It holds the eighteen laws, the eleven passes, the twelve engines, the workflow and the deterministic verifier. Read it first. This bundle is what that skill reaches into.
| When you are | Read |
|---|---|
| Writing in a specific genre | references/genre_packs/ — pick the pack, honor its arena, its objection set and its structural contract |
| Writing in someone's voice | references/voiceprints/ — a voiceprint is measured from real text, never guessed from adjectives |
| Diagnosing a flat or machine-sounding draft | references/anti_patterns/ then references/positive_patterns/ |
| Running Pass 3 — the diagnostic scan | references/diagnostics/ |
| Executing or extending any of the eleven passes | references/compiler/ |
| Grounding a claim about writing research | references/academic/ |
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.
- 4d ago First seen · 75 lines · 185 tokens per session scan A 15f54ef8a320
writing-intelligence-craft is a skill published in the GitHub repository antonio0720/writing-intelligence (13 stars, last pushed 26d ago), licensed MIT. It adds 185 tokens to every session and 1,214 once invoked, about $0.0009 per session on Opus 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.
Other skills, from other repositories
novel-original-writing
中文名：原创剧本流。原创剧本流协议（原创入口，与 02 改编流并行）。触发词：原创剧本/概念到剧本/从零写剧本/原创短剧/原创概念/无原著改编/剧本流。概念直接到剧本，不经过小说：七阶段（种子→世界→人物→结构→载体→表演→余震）从概念长成可拍剧本，质量门 20 条 + 导演模式驱动。.
novel-to-screen-adaptation
Novel-to-Screen Adaptation Protocol. Not a script-writing skill. A translation machine that decomposes a novel into proposition, engine, objects, scenes, episode structure, and screenplay — in that order. 中文名：改编流 / 小说改编剧本协议。触发词：改编流/小说改编/剧本改编/改编剧本/小说转剧本/小说到剧本/改编协议。将小说分解为命题、发动机、物件、场景、剧集结构与剧本的转译协议（不是写剧本的 skill）。.
video-prompt-adapter
表演驱动的视频提示词。中文名：表演层视频提示词 / 07 表演层。触发词：视频提示词/表演提示词/提示词生成/表演层/镜头提示词/短剧提示词。三层结构：表演层（角色内在/表演心理学/微表演/否定约束/风格体系/镜头-关系）+ 平台能力层（Seedance 2.5 公式/@引用/时间戳/参数）+ 管线接口（06 分镜包/04 知识库/continuity-locks/导演模式）。.
novel-review
小说理论评审：按《人物篇》《命运篇》《文学天演论》《故事天演论》标准诊断小说问题出在哪 + 文笔判断（现场八查）+ 天演打分（W 值）。触发词：小说评审/小说诊断/review 小说/看看这部小说/问题出在哪/理论评审/给这部小说打分/文笔怎么样/这段写得如何。三层扫描（人物/结构/结局）+ 文笔扫描（现场八查）+ 红黄牌分级 + 证据引用 + 天演评分（继承神之公式结构，判据接当前理论），每条结论可追溯到理论条目。.
skill-evolution
技能进化复盘：从与用户的一轮/一段对话中学习用户风格（纠正信号/风格信号/新约定），生成风格档案，并对相关 skills 提出自主优化。触发词：技能进化/复盘技能/优化技能/学我风格/风格学习/技能复盘/这一轮学到了什么。与 retrospective 的边界：retro 是调参循环的 W 变量复盘；本 skill 是风格学习 + 技能自我进化。.
skill-e2e-test
技能端到端测试：让 AI 按 target skill 的 Procedure 全流程跑一遍（模拟导演 + 机械扫描），逐阶段对照协议找问题，输出 E2E 报告喂给 skill-evolution。触发词：技能端到端测试/E2E 测试/跑一遍 skill/技能验收/全流程测试/测试技能。与 e2e-summary 的边界：e2e-summary 汇总代码测试结果；本 skill 是协议级 E2E（AI 执行创作流程 + 协议符合性检查）。.