Borrowing it
Nothing to install: this file belongs to xiaojilele-glitch/WhyBuddy. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/xiaojilele-glitch/WhyBuddy/main/CLAUDE.mdgit clone --depth 1 https://github.com/xiaojilele-glitch/WhyBuddyWrote 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/xiaojilele-glitch/whybuddy/claude-md)<a href="https://agentmods.dev/instructions/xiaojilele-glitch/whybuddy/claude-md"><img src="https://agentmods.dev/badge/instructions/xiaojilele-glitch/whybuddy/claude-md/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/instructions/xiaojilele-glitch/whybuddy/claude-md"><img src="https://agentmods.dev/badge/instructions/xiaojilele-glitch/whybuddy/claude-md.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 | $0.03550 | $0.03550 |
| Opus 5.5 | $0.01420 | $0.01420 |
| Sonnet 5.5 | $0.00710 | $0.00710 |
| Haiku 4.5 | $0.00355 | $0.00355 |
Grade A, and why
WhyBuddy 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 1mo 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 — 220 lines — stays where its author put it; the contents beside it link to each section on GitHub.
给接手这个仓的人(含 AI)
这个仓的知识不在文档里,在代码注释和测试 docstring 里。随便打开一个
services/*.py,模块头往往是几百字的事故记录:哪天真机踩了什么、为什么这么写、
上一版错在哪。那不是话多,是这个仓的记忆方式——读代码前先读它的注释。
下面是从那些注释里反复出现的纪律,按被违反的频率排。
一、动手之前,先确认哪条链真的在跑
这是本仓最贵的一条。 2026-08-16 一天之内因为违反它打偏三次:
改了闭环重建那一步 而模型是主循环里生成的 (v5_full_driver:1260 vs :531)
改了提示词的收尾 同样在没被使用的那一步
把合并接在老生成器上 真正在跑的是 spec-first (v5_capability_executor:492)
三次的代码本身都是对的,三次都因为装在不通电的插座上而毫无效果,三次都靠真机 日志才发现——每次损失 25 分钟真机跑 + 一次部署。
做法:改之前先在目标路径打一行日志或加一条断言,确认它真的被执行到。
tests/test_refine_merge_reaches_the_live_path.py 就是这条纪律的具象化。
一之二、护栏装对了地方,条件却永远不成立
这是第一条的变种,2026-09-04 一夜里发作两次,同一个人写的两条护栏。 两次都不是装错位置——装在真跑的那条路上,但它依赖的输入在真机上根本不出现:
护栏 「文本只许压残留 hop,不许压显式意图」 写成 forced is None or forced in FACTORY_HOPS
真机 首轮点火按设计不传 forcedTool → forced is None → 文本照样赢
[control] forced hop=structure hasSpec=0 hasPages=0 后面连 capabilityPlan= 都没有
护栏 「已有 SPEC、host 没点 spec,不许把 spec 塞回去」 加在 _factory_tools_from_state
真机 run_spec_first 转手 select_workflow(tools=…) 拿 preset.tools 当计划,
['pages','structure','bind'] → ['spec','pages','structure','bind'] spec 又回来了
stage=specfirst.spec ms=18428 …… 页面落库:0 份
两次的单测都是绿的,因为判据自己构造了护栏需要的那个输入(喂
{'forcedTool':'rehearse'}),而真机不喂。第二次更贵:用户答完假设卡点确认,
整跳预算烧在重起草一份不一样的 SPEC 上(5 页 16 节点 → 6 页 17 节点),
页面一页没画,而刚确认的假设是针对旧那份的。
做法:护栏的判据必须喂真机那一发的原样载荷,不许自己拼一个。
拿不到就先去日志里抄一行。写完再问一句:这个条件在真机上真的会成立吗?
tests/test_topic_is_not_an_instruction.py 与
tests/test_remaining_chain_does_not_redraft_spec.py 是这条的具象化——
前者第一条判据先证明「不设边界时文本确实会赢」,后者干脆直接执行产线源码,
不重抄逻辑(重抄的判据只能证明"我抄对了")。
二、判据必须能被变异咬住
写完测试,把修复改回去,确认它变红。没红就是判据没用。
本仓被这条抓到过的真实形态:
- 判据 grep 源码里的标识符,而那个词同时出现在文档字符串里 → 变异后照样绿。 修法:匹配前先剥注释。
- 判据写
"Produce the complete" not in tail,而实际收尾是"Produce the five-system JSON now."(不含 complete)→ 断言直接打空。 修法:盯语义(还在要求"产出一份"),别盯某句话的字面。 - 11 条测试全绿,但把调用点删掉照样全绿——它们只直接调那个函数,从没验证 它接在链路上。这就是下面第三条。
三、"闸全绿但东西没了"
本仓数到第十次以上的失败形态:正向判据齐全,反向判据缺失。
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.
- 1mo ago First seen · 220 lines · 3,550 tokens per session scan A ca0bd4cd86a7
WhyBuddy CLAUDE.md is an instructions file published in the GitHub repository xiaojilele-glitch/WhyBuddy (380 stars, last pushed 1mo ago), licensed MIT. It adds 3,550 tokens to every session, about $0.0142 per session on Opus 5.5. 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 2026-09-23.
WhyBuddy CLAUDE.md compared with instructions from other repositories
| Mod | Type | Stars | Last push | Licence | Scan | Tokens | Works with |
|---|---|---|---|---|---|---|---|
| WhyBuddy CLAUDE.mdxiaojilele-glitch/WhyBuddy | Instructions file | 380 | 1mo ago | MIT | A | 3,550 | Claude Code |
| vscode buildNext.instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 9,719 | GitHub Copilot |
| next.js AGENTS.mdvercel/next.js | Instructions file | 143k | today | MIT | A | 7,461 | Codex, OpenCode |
| langchain AGENTS.mdlangchain-ai/langchain | Instructions file | 149k | today | MIT | A | 4,680 | Codex, OpenCode |
| vscode oss-third-party-notices.instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 5,516 | GitHub Copilot |
| vscode copilot-instructions.mdmicrosoft/vscode | Instructions file | 195k | today | MIT | A | 3,093 | GitHub Copilot |
| deepseek-harness AGENTS.mddeepseek-ai/deepseek-harness | Instructions file | 274k | 2d ago | MIT | A | 4,313 | Codex, OpenCode |