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 seed-forge/harness-ai-kit --skill work-sc-document-sop-buildergit clone --depth 1 https://github.com/seed-forge/harness-ai-kitWrote 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/seed-forge/harness-ai-kit/work-sc-document-sop-builder)<a href="https://agentmods.dev/skills/seed-forge/harness-ai-kit/work-sc-document-sop-builder"><img src="https://agentmods.dev/badge/skills/seed-forge/harness-ai-kit/work-sc-document-sop-builder.svg" alt="Measured on agentmods" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00080 | $0.01875 |
| Opus 5 | $0.00040 | $0.00937 |
| Sonnet 5 | $0.00016 | $0.00375 |
| Haiku 4.5 | $0.00008 | $0.00187 |
Grade A, and why
work-sc-document-sop-builder 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 8d 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 — 219 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Document Reference SOP Builder
将参考文档中的隐含规则整理成可复用的文档编制 SOP、格式规范和导出模板能力。
适用场景
- 用户给出一个
docx、md、参考底稿或历史成品，希望后续类似文档都按它来编 - 需要从参考文件中提取章节结构、内容要求、固定口径、版式规范和导出模板
- 需要把一次性的“照着参考文件做”沉淀成项目级 Skill 或 SOP
不负责的内容
- 不替代具体业务领域的专业写作技能
- 不自动判断法律、合规或行业规范是否正确
- 不直接完成所有文档编写任务，而是先把规则体系抽出来
总体目标
把参考文件中的规则拆成以下几层：
- 内容层规则
- 结构层规则
- 表达层规则
- 版式层规则
- 导出层规则
- 工作流层规则
最终形成：
- 可复用 SOP
- 可复用导出模板
- 可复用项目级 Skill
- 可引用的规则说明文档
标准流程
SOP 1：确认参考源
先确认参考材料类型：
docxmdpdf- 已有 Skill / 已有模板说明文档
若参考源是 docx，优先使用 markitdown 一类转换能力转成 md 再分析。
SOP 2：拆分规则层次
阅读参考文档时，不只看“写了什么”，而要拆成以下几类规则：
-
结构规则
- 有哪些章节
- 章节顺序是否固定
- 是否存在常用子章节
-
内容规则
- 每一章要求写什么
- 哪些内容必须展开
- 哪些内容只是补充说明
-
表达规则
- 是公文语气、技术语气、汇报语气还是宣传语气
- 是否存在固定称呼、固定衔接语、固定标题口径
- 哪些表达应避免
-
版式规则
- 字体、字号、行距、缩进、页边距
- 首页信息栏、表格、页眉页脚、目录等
-
导出规则
- 输出为
docx、pdf、html还是其他格式 - 是否区分不同 profile
- 文件名与正文标题区的版本信息如何分离
- 输出为
-
工作流规则
- 永远先改
md再导出 - 先做结构确认还是先做版式确认
- 内容确认后才进入导出
- 永远先改
SOP 3：抽取显性要求与隐性要求
区分两类信息：
- 显性要求：文档中直接写明的章节要求、格式要求、填写说明
- 隐性要求：从成品排版、措辞、结构重复中推断出的稳定规则
抽取时应明确标记：
- 哪些来自原文直接要求
- 哪些属于分析归纳
SOP 4：沉淀为规则文档
建议至少产出以下一种或多种资产：
章节编写要求.md导出模板与格式要求.md编制工作流.md- 项目级
SKILL.md
SOP 5：封装为项目级 Skill
当以下条件满足时，建议正式沉淀为项目级 Skill：
- 该类文档未来还会重复编制
- 结构和格式规则较稳定
- 用户已明确要求长期复用
- 已经完成至少一次真实项目验证
Skill 内应明确：
- 适用文档类型
- 输入材料
- 结构规则
- 内容规则
- 版式规则
- 导出规则
- 与其他专业写作 Skill 的协同关系
SOP 6：接入知识管理
该 Skill 一旦形成，应同步：
- 建立或更新
Skill卡片 - 回填
Skill收藏库 - 标注它是通用编制方法、项目级执行资产，还是可供其他文档类型参照的基础 Skill
使用原则
原则 1：先抽规则，再套用
不要直接把参考文档机械改名后复用。先提取规则，再决定是形成新模板、形成新 Skill，还是补充到旧 Skill 中。
原则 2：内容规则与版式规则分开沉淀
同一参考文档里，内容要求和导出样式可能分别适用于不同场景，应分开维护。
原则 3：先 md，后导出
即使目标是 Word 或 PDF，也优先把结构和规则沉淀到 md。
原则 4：优先形成 profile
若存在多种交付样式，应明确形成多个 profile，例如：
client-reviewinternal-reviewgov-stylereference-template
推荐输出结构
# 一、参考文档来源
# 二、章节结构要求
# 三、章节内容要求
# 四、语言与表达要求
# 五、版式与导出要求
# 六、工作流与使用边界
与其他 Skill / 能力的关系
What ships with it
7 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.
- 8d ago First seen · 219 lines · 80 tokens per session scan A 031fa53b3cd4
work-sc-document-sop-builder is a skill published in the GitHub repository seed-forge/harness-ai-kit (22 stars, last pushed 7d ago), licensed Apache-2.0. It adds 80 tokens to every session and 1,875 once invoked, about $0.0004 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
secrets-in-file-metadata
Extract and interpret embedded file metadata with exiftool — EXIF GPS coordinates, camera make, model and serial, DateTimeOriginal and CreateDate timestamps, XMP and IPTC fields, and Office and PDF properties such as Author, Company, LastModifiedBy, template paths and revision counts. Use when reading EXIF from a…
audit-bundle-size
Analyse and shrink JavaScript bundle size for any web app. Use when asked to "reduce bundle size", "analyse bundle", "tree shaking", "lazy loading", "code splitting", "slow initial load", "large JS", "chunk size", "build performance", "LCP caused by JS", "why is the bundle so big", or "first load JS too large".
RAG Chunking Strategy Advisor
Given a document type and retrieval goal, recommends the optimal chunking strategy for a RAG pipeline to minimize retrieval failures.
pydicom
Use pydicom to read, inspect, write, transform, and safely preflight local DICOM datasets and pixel data. Applies to DICOM metadata, transfer syntaxes, compression plugins, frames, private elements, JSON, and bounded de-identification review.
baoyu-youtube-transcript
A tool for downloading the written captions, subtitles, chapter information, speaker labels, and cover image from a YouTube video using its URL or ID.
comprehensive-enrichment
Enrich any person or company from any identifier — email, name, LinkedIn URL, domain, company name, Twitter/X handle. Use when asked to enrich, look up, or research a lead, contact, person, or company.