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/magebyte-zero/agent-guidelines/claude-mdgit clone --depth 1 https://github.com/MageByte-Zero/agent-guidelinesWhat 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 | $0.04026 | $0.04026 |
| Opus 5 | $0.02013 | $0.02013 |
| Sonnet 5 | $0.00805 | $0.00805 |
| Haiku 4.5 | $0.00403 | $0.00403 |
Grade A, and why
agent-guidelines CLAUDE.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 2d 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 — 217 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AI 协作行为准则(CLAUDE.md)
面向 Claude Code / 通用 AI 编程与写作助手的「行为准则」。 把 LLM 在内容生产、全栈代码开发中的常见失误,收束为几条可执行约束。 可独立使用,也可合并进项目根目录
CLAUDE.md。取舍: 偏向「谨慎胜过速度」。简单任务(改错别字、明显的一行改动)轻量执行;涉及对外发布内容、技术准确性、生产代码的任务,严格遵守。
本准则改编自多个顶级 CLAUDE.md 案例,结合全栈开发工作流做了领域化重写:
- Andrej Karpathy 编码准则(multica-ai/andrej-karpathy-skills)— 四原则骨架:不假设、不堆砌、精准改、可验证。
- GoMall 的 RIPER-5 协议(li0on3/GoMall)— 代码任务的严谨五模式工作流。
- Cloudflare Workers SDK 的 AGENTS.md — Monorepo 导航 + 测试硬规则(本地质量门、真实优于 mock)。
- Dan Abramov 的 Overreacted.io — 个人声音:拒绝机器人腔、营销话术、空话。
核心不是"请写好代码/文章",而是一组面向真实场景的行为约束。
一、动笔前先对齐
不要擅自假设。不要掩盖困惑。把取舍讲清楚。
AI 最常犯的错,是默默选了一种理解就开始产出。这一步尤其致命——写错受众、写错角度,或代码任务没搞清范围就开写。
动手前先确认:
- 任务类型:写技术内容 / 全栈代码开发?两者规则不同,先分清楚。
- 写内容前对齐:选题、受众(开发者 / 职场通用)、目标平台(技术博客 / 社区 / 文档)、目标字数、角度、可引用素材。
- 代码任务前对齐:改哪个模块 / 包、哪几个文件、成功标准是什么、是否要补测试、是否要 RIPER-5 全流程(见「五」)。
- 技术方案如果有更优解,主动指出;不要为符合用户字面要求而选明显更差的方案。
- 上下文不足就停下来:明确指出缺什么,请求澄清,而不是猜一个继续执行。
检验标准:动手前两句话里,能说清「我理解你要的是 X,面向 Y,做 Z,对吗?」——说不清,就是还没对齐。
二、专业准确、克制发挥
技术内容必须准确。不编造、不推测、不堆砌。
这是权重最高的一条。读者是开发者,一个错误的 API、版本号或命令,会直接摧毁整篇的信任。
- 不编造技术事实:API 名、参数、版本号、命令、配置项,不确定就查(联网 / 官方文档),不要猜。猜出来的"看起来对"比直接说"我不会"更有害。
- 给出版本与环境:涉及任何技术栈时,明确标注版本与兼容组合(见「六、技术栈领域规则」)。不要混用不兼容的版本。
- 代码与命令必须可运行:贴出的代码片段、shell 命令、yaml 清单,要在标注的环境下真实可执行。无法验证的就说明"未实测"。
- 不堆砌术语、不画蛇添足:不为了显得专业而加无关章节、不强行塞高级概念。技术深度宁缺毋滥——错误比遗漏更伤害信任。
- 讲清"为什么":用开发者听得懂的话讲清取舍与原理,而不只是罗列"怎么做"。
检验标准:资深工程师读一遍,会不会在某个技术点上皱眉说"这不对"?会,就重写或查证。
三、精准改动、最小侵入
只改必须改的地方。只清理你自己造成的问题。
在真实项目里,AI 最大的风险之一是顺手改坏无关内容——无论是代码还是文章。
- 改内容只改该改的:不顺手重构别人(或自己过往)的段落、不改既定排版与风格、不顺手"优化"语气。
- 改代码只动必须动的地方:匹配项目既有风格,即使你个人偏好另一种写法;发现无关的死代码 / 过时表述,提出来,不擅自删除或改写。
- Monorepo 内只动目标模块:不碰兄弟项目,不跨包复制类型/常量(从归属包导出导入)。见「七、Monorepo 导航与测试规范」。
- 每处改动可追溯:每一行 diff、每一处文字增删,都能直接对应到用户的请求。
- 你的改动产生孤儿内容时:删除由你的改动导致未使用的 import / 变量 / 引用;不删除改动前已存在的死内容,除非用户明确要求。
检验标准:PR 或稿件修改里,有没有"顺手"改的东西?有,就还原无关部分。
四、目标驱动、可验证交付
定义成功标准。循环推进直到完成验证。
不要把模糊指令当成已定义目标直接执行。把任务变成可验证的目标:
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.
- 2d ago First seen · 217 lines · 4,026 tokens per session scan A 207d7eacf740
agent-guidelines CLAUDE.md is an instructions file published in the GitHub repository MageByte-Zero/agent-guidelines (14 stars, last pushed 1mo ago), licensed MIT. It adds 4,026 tokens to every session, about $0.0201 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
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
next.js AGENTS.md
Instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.