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 instructions/peiyucn/epytor/agents-mdgit clone --depth 1 https://github.com/peiyucn/epytorWrote 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/instructions/peiyucn/epytor/agents-md)<a href="https://agentmods.dev/instructions/peiyucn/epytor/agents-md"><img src="https://agentmods.dev/badge/instructions/peiyucn/epytor/agents-md.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 | $0.04047 | $0.04047 |
| Opus 5 | $0.02024 | $0.02024 |
| Sonnet 5 | $0.00809 | $0.00809 |
| Haiku 4.5 | $0.00405 | $0.00405 |
Grade A, and why
epytor AGENTS.md 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 today.
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 — 316 lines — stays where its author put it; the contents beside it link to each section on GitHub.
项目指令 — epytor
语言
- 始终用简体中文回复
需求
- 新功能设计文档放在
docs/specs/,文件名YYYY-MM-DD-<功能名>.md - 先写 spec 再开发——明确需求范围、交互边界、验收标准
- 配置项参考见开发 → 配置参考
开发
基本规则
- 包管理器:必须用
pnpm,禁止 npm/yarn - 构建:修改代码后执行
pnpm build验证编译无误 - 调试:F5 启动扩展调试实例(
.vscode/launch.json) - 语言:全部 TypeScript;Extension 端用
tsconfig.json,WebView 端用tsconfig.webview.json - 双目标构建:
dist/extension.js(Node.js)+dist/webview.js(Browser),由esbuild.mjs完成 - Git commit 规范:commit 描述用中文,类型前缀保留英文(
feat:、fix:、refactor:、chore:、docs:等)。例:feat: 新增XXXX功能、fix: 修复XXXX问题 - 逐项提交:todo list 中每完成一个独立任务必须单独
git commit,禁止多个 todo 混在一个 commit 中(方便出问题时精确回溯) - 诚实原则:不确定的事直接说"不确定",禁止编造 URL、issue 编号、API 接口、文档引用或任何事实性信息
- 优雅原则:禁止 hack 或补丁式写法,优先使用框架/库官方 API、CSS 变量、配置回调等正路方案
- 自检原则:代码移动/提取后必须搜索确认旧位置已删除,不得留有死代码或同名遮蔽;标记 roadmap 条目完成前逐项列出实际完成项与未完成项,不得将部分完成标记为整体完成
- 查证原则:引用文件位置、函数名、调用关系时,若不确定则先 grep 确认再写,禁止凭记忆编造
架构约束
- WebView ↔ Extension 通信只通过
webview/messaging.ts中封装的函数 - WebView 侧不直接
importVSCode API,通过acquireVsCodeApi()获取句柄 - CSS 必须使用
--vscode-*变量以适配亮/暗主题 - 不在模块外部维护全局状态(单例除外,如 editor view)
关键文件速查
src/extension.ts — 扩展入口,注册 CustomEditorProvider
src/MarkdownEditorProvider.ts — Provider 核心(消息路由、自动保存、revert)
src/utils/getNonce.ts — CSP nonce 生成
src/utils/imageService.ts — 图片本地保存(MD5 去重)+ 服务器上传
src/i18n/webviewTranslations.ts — WebView 翻译数据
webview/index.ts — WebView 入口(消息路由、DOM 事件委托、品牌标识注入)
webview/editor.ts — CrepeBuilder 入口(Milkdown 7.21.2 + Crepe 原生功能注册)
webview/messaging.ts — WebView ↔ Extension 消息协议(唯一通信层)
webview/style.css — VSCode 主题全覆盖(--vscode-* CSS 变量,覆盖 Crepe 组件)
webview/i18n/index.ts — t() / kbd() 翻译函数
webview/ui/icons.ts — SVG 图标
webview/ui/tooltip.ts — Tooltip 组件
webview/utils/themeBus.ts — Mermaid/CodeMirror 深浅主题统一事件总线
webview/components/selectionToolbar/index.ts — 选区变更回调(驱动源码行号映射)
webview/components/toc/index.ts — 目录(TOC)面板(吸底工具栏下方、可固定、可拖拽宽度)
webview/components/imageView/index.ts — 图片 NodeView(选中/lightbox/工具栏/缩放 handle)
webview/components/findBar/index.ts — 编辑器内查找栏(Cmd/Ctrl+F)
webview/components/pathLink/ — 路径链接自动补全
webview/headingIds.ts — 标题 id 管理(不操作 DOM,仅保留签名)
docs/specs/ — 功能 spec 文档
docs/roadmap.md — 项目路线图(面向用户的功能规划)
docs/tech-debt.md — 技术债务清单(面向开发者的代码改进)
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.
- today Changed · +2 lines · +163 tokens per session 2e3e8d5c2563
- 4d ago First seen · 314 lines · 3,884 tokens per session scan A 0266113dddfd
epytor AGENTS.md is an instructions file published in the GitHub repository peiyucn/epytor (12 stars, last pushed 2d ago), licensed MIT. It adds 4,047 tokens to every session, about $0.0202 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 instructions, from other repositories
kilocode AGENTS.md
AGENTS.md instructions for Kilo-Org/kilocode, covering agents.md, build and dev, quality checks, products and package instructions.
awesome-vscode CLAUDE.md
Instructions for viatsko/awesome-vscode, covering claude.md, what this repo is, adding an entry, themes and linting / ci.
vscode-peacock AGENTS.md
AGENTS.md instructions for johnpapa/vscode-peacock, covering peacock — agent guide, project overview, repository structure, tech stack and build & run.
vscode-peacock copilot-instructions.md
Copilot instructions for johnpapa/vscode-peacock, covering copilot instructions — peacock, project type, typescript conventions, vs code extension patterns and color handling.
language-tools AGENTS.md
AGENTS.md instructions for prisma/language-tools, covering agent guide for prisma language tools, project overview, packages, documentation and quick start.
getspecstory copilot-instructions.md
Copilot instructions for specstoryai/getspecstory: Examine the file changes for the pull request.