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.
git clone --depth 1 https://github.com/jzl-maker/EA-SKILLWrote 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/commands/jzl-maker/ea-skill/doc)<a href="https://agentmods.dev/commands/jzl-maker/ea-skill/doc"><img src="https://agentmods.dev/badge/commands/jzl-maker/ea-skill/doc/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/commands/jzl-maker/ea-skill/doc"><img src="https://agentmods.dev/badge/commands/jzl-maker/ea-skill/doc.svg" alt="Reviewed on agentmods" width="80" 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 5.1 | $0.00000 | $0.02229 |
| Opus 5 | $0.00000 | $0.01115 |
| Sonnet 5 | $0.00000 | $0.00446 |
| Haiku 4.5 | $0.00000 | $0.00223 |
Grade A, and why
doc 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 yesterday.
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 — 143 lines — stays where its author put it; the contents beside it link to each section on GitHub.
命令: /ea doc (文档识别)
功能
把 PDF / Word(.docx) / Excel(.xlsx) 转成 Markdown 或 JSON，供 AI 直接消费。
用于识别：芯片手册/数据手册、需求规格文档、引脚分配表、寄存器表、测试报告、协议说明。
只做解析，不做 OCR。扫描件（无文本层的 PDF）明确报错，不做图片文字识别。
触发
/ea doc <路径> # Markdown（默认，表格超 20 行截断）
/ea doc <路径> --outline # 只输出结构大纲（快速判断文档讲什么）
/ea doc <路径> --full # 不截断
/ea doc <路径> --tables # 只输出表格
/ea doc <路径> --pages 1-20 # PDF 页范围（也支持 1,5,7-9）
/ea doc <路径> --sheet Sheet1 # 只解析指定工作表（可重复）
/ea doc <路径> --json # 结构化 JSON
/ea doc <路径> --out <文件> # 同时写入文件
/ea doc <路径> --keep-watermark # 保留平铺水印行（默认自动过滤）
/ea doc --scan <目录> # 列出目录下可识别文档
执行流程
- 【依赖检查】 首次运行前确认
pdfplumber（PDF 需要；DOCX/XLSX 零依赖）：py ~/.claude/skills/EA-SKILL/tools/doc-reader/scripts/deps_check.py --detect - 【路径确认】 确认目标文件存在、扩展名受支持
- 【调用工具】
py ~/.claude/skills/EA-SKILL/tools/doc-reader/scripts/doc_reader.py <路径> [参数] - 【解读】 按用户意图消费结果（见下方「串联场景」）
支持矩阵
| 格式 | 解析方式 | 依赖 | 说明 |
|---|---|---|---|
pdfplumber |
pip | 需文本层；含表格提取（--pages 分段） |
|
Word .docx |
zipfile + XML |
无 | 保留标题层级 + 表格 |
Excel .xlsx |
zipfile + XML |
无 | 多工作表 + sharedStrings + 内联字符串 |
不支持（明确报错并给替代方案）：
| 格式 | 提示 |
|---|---|
| 扫描件 PDF | 无文本层 → 改用 Read 工具走视觉识别 |
.doc / .xls |
老二进制格式 → 另存为 .docx / .xlsx |
.pptx / .ppt |
另存为 PDF 后再 /ea doc |
.csv / .txt / .md |
直接用 Read 工具读取 |
输出格式
# <文件名>
> 来源: `<路径>`
> 格式: pdf | 解析: 2026-09-10 09:45
> pages_total=120 | pages_parsed=120 | pages_with_text=118 | tables=3
> ⚠️ 2 页无文本层（可能是扫描页/纯图页），已跳过其文本
## 第 1 页
<正文>
| 列1 | 列2 |
|---|---|
- 标题行
>元信息 + 警告，便于 AI 判断可信度（有几页没解析到） --json输出{path, format, meta, warnings, sections[]}，sections[].kind ∈ heading/paragraph/page/table/sheet
平铺水印过滤（默认开启）
供应商给的规格书常带斜向铺满页面的水印（如"仅供东屋参考"）。这类水印有文本层， 会被逐字提取成大量重复的单字行（实测可占非空行的 80%）。后果：
--outline的每页 200 字摘录被水印占满，"快速判断文档讲什么"的作用完全失效- 表格数值被粘上水印字：
| 参 57600 |、| 屋 Φ20.90 |、| 供 V1.0.5 |
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.
- yesterday Changed · +33 lines 46c493879c74
- 2d ago First seen · 110 lines · 0 tokens per session scan A 27d5f36014b8
doc is a command published in the GitHub repository jzl-maker/EA-SKILL (6 stars, last pushed yesterday), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 2,229 tokens. 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 2026-09-10.
Other commands, from other repositories
audit-xls
Audit an Excel workbook — formula errors, hardcoded cells, calculation arc cross-validation.
generate-pdf-document
Specify a generated PDF that survives variable content, with fonts, page breaks, and accessibility handled.
convert
Use when the user attaches a bank statement PDF or asks for CSV/XLSX/QBO/Xero export of a converted statement. Do not use for spending analysis, reconciliation, or benchmark scoring — those have their own skills.
batch-fill
Fill a form once per row of a CSV with Emboss.
extract
Extract product information from screenshots, PDFs, and catalog images in for-ai/ into a structured CSV at data/extracted/products.csv.
ars-format-convert
ARS academic-paper format-convert mode — convert to LaTeX / DOCX / PDF / Markdown.