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/koobai/blog/agents-mdgit clone --depth 1 https://github.com/koobai/blogWrote 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/koobai/blog/agents-md)<a href="https://agentmods.dev/instructions/koobai/blog/agents-md"><img src="https://agentmods.dev/badge/instructions/koobai/blog/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 | $0.03953 | $0.03953 |
| Opus 5 | $0.01976 | $0.01976 |
| Sonnet 5 | $0.00791 | $0.00791 |
| Haiku 4.5 | $0.00395 | $0.00395 |
Grade A, and why
blog 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 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 — 231 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
本文件是 AI 编程助手与自动化代理修改惊蛰项目时必须遵守的项目契约。开始任何修改前，应完整阅读本文件以及与任务直接相关的 docs/ 文档。
项目目标
惊蛰是 Koobai 的真实生产博客，也是开放给他人使用的 Hugo 个人数字生活发布系统。
项目整理遵循最高优先级原则：
开源能力只做增量建设；Koobai 现有的内容、发布入口、数据同步、自动化和线上服务默认全部保持向后兼容。
当前状态
仓库同时承载 Koobai Production 和可复用的 Core/可选服务源码。当前生产代码包含 Koobai 专属配置；Core 初始化、统一检查和五个 Worker 开源部署包已经可用。任何 GitHub、Cloudflare 或生产服务的外部写操作仍需项目所有者单独授权。
源文件地图
config/_default/：所有环境共享的 Hugo 结构和 Core 默认参数。config/production/：Koobai 生产身份、菜单、功能开关和公开服务配置。config/development/：保持hugo server与生产参考站一致的本地覆盖。data/jingzhe/features.json：机器可读功能注册表与数据职责目录；每份公开/私有 JSON 都声明写入者、读取者、生成属性和 Schema。data/jingzhe/exercise.json：运动名称、颜色、分组和食物换算的跨语言单一来源。data/jingzhe/zouguo_boundary_catalog.json：走过构建期完整行政区边界目录；浏览器只接收当前内容引用的裁剪子集，并须保留上游许可证。data/exercise/activities.json：来源无关的原始运动事实；只由同步网关写入，只由处理器读取。jingzhe/exercise_contract.py：Python 运动契约加载器，不得复制一套常量。schemas/：Front Matter、站点参数与公开 JSON Schema。archetypes/：新文章和新唠叨模板。content/posts/：长篇随笔。content/laodao/YYYY/MM/：短动态。content/zouguo/：独立走过 Markdown；带“走过”Tag 的唠叨和随笔仍以各自原文件为来源。content/pages/：独立页面和管理入口。themes/jingzhe_v3/layouts/：Hugo 全局与内容类型模板；layouts/pages/只放content/pages/的专用页面布局。themes/jingzhe_v3/assets/：主题原生 CSS、JavaScript 与 Hugo Pipes 资源。themes/jingzhe_v3/assets/css/style.css：唯一全站样式入口；运动和走过分别通过各自目录中的index.css组合职责分片。themes/jingzhe_v3/assets/js/navigation/：灯箱、桌面导航、移动导航和共享滚动计算；构建时仍合并为js/site.js。themes/jingzhe_v3/assets/js/pages/：普通页面交互脚本，经 Hugo Pipes 指纹化输出。themes/jingzhe_v3/assets/js/zouguo/、layouts/zouguo.html与_partials/zouguo/：走过地图、列表、统一 feed 和构建期边界裁剪。themes/jingzhe_v3/assets/js/exercise/：运动数据模型、日历 UI、隐私路线、Mapbox、海报与兼容控制器；构建时合并为一个脚本。themes/jingzhe_v3/assets/js/pages/editor-core.js：两个在线编辑器共享的鉴权、草稿、标签、上传、预览与 GitHub 原语。themes/jingzhe_v3/assets/js/pages/editor-laodao.js、editor-post.js：两个写作页面各自的 UI 与发布行为。static/js/：按上游许可证原样保留的第三方浏览器脚本。assets/data/movies.json：公开观影数据；由sync_movies.py原子更新。assets/data/exercise/：处理后运动、公共地标路线和月报等公开数据；App 与同步网关不得直接读写处理后的activities.json。process_activities.py、monthly_coach.py：Actions 和旧调用方保持不变的兼容入口。jingzhe/activity_processing.py、public_routes.py：运动数据加工、展示字段和隐私公共路线。jingzhe/monthly_stats.py、monthly_reports.py：月报统计证据、模型调用、校验与冻结状态机。sync_movies.py：豆瓣观影数据同步。tests/：Python、浏览器脚本、Worker、文档和隐私契约测试。workers/：Publisher、Drafts、Comments、Likes、Activity Sync 五个独立 Cloudflare Worker，以及示例配置、D1 迁移和 OpenAPI。.github/workflows/：同步、处理、构建与部署流程。docs/：安装、架构、功能、配置、隐私和 AI 协作文档。
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 · 231 lines · 3,953 tokens per session scan A 5da98e2ab9eb
blog AGENTS.md is an instructions file published in the GitHub repository koobai/blog (54 stars, last pushed yesterday), licensed MIT. It adds 3,953 tokens to every session, about $0.0198 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
blowfish CLAUDE.md
Claude Code instructions for nunocoracao/blowfish, covering blowfish, working on the theme and agent skill — keep it in sync.
blog CLAUDE.md
Claude Code instructions for gbmhunter/blog, covering units in latex/katex equations, multiplication in latex/katex equations, katex gotchas (beyond units/multiplication above), images and diagrams and mdx components.
vercel-action CLAUDE.md
Instructions for amondnet/vercel-action, covering claude.md, project overview, package manager, essential commands and development.
freek.dev CLAUDE.md
Claude Code instructions for spatie/freek.dev, covering laravel boost guidelines, foundational context, skills activation, conventions and verification scripts.
setup-chromedriver CLAUDE.md
Instructions for nanasess/setup-chromedriver, covering claude.md, commands, build and development, running a single test and or.
databricks-template CLAUDE.md
Instructions for andre-salvati/databricks-template, covering claude.md, project overview, tooling: mcp servers, cli, skills → see specs/tooling.md, commands and architecture & data model → see specs/.