Borrowing it
Nothing to install: this file belongs to Cyrus580529/Attest. 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/Cyrus580529/Attest/master/CLAUDE.mdgit clone --depth 1 https://github.com/Cyrus580529/AttestWrote 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/cyrus580529/attest/claude-md)<a href="https://agentmods.dev/instructions/cyrus580529/attest/claude-md"><img src="https://agentmods.dev/badge/instructions/cyrus580529/attest/claude-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 5.1 | $0.13350 | $0.13350 |
| Opus 5 | $0.06675 | $0.06675 |
| Sonnet 5 | $0.02670 | $0.02670 |
| Haiku 4.5 | $0.01335 | $0.01335 |
Grade A, and why
Attest 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 8d 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 — 212 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Attest 开发规则
可信、自证的网页 agent 内核。从 SkillFlow/FlowOps 干净重写而来。 本文件是这个仓库的开发地基——只保留真正用得上的要领,砍掉重复冗杂。
一、架构红线(永不违反)
这些不变量定义了 Attest,是它区别于"抓取式 agent"的根本。任何改动都不得破坏:
- 模型只提议,harness 校验后才执行。模型给的
ref必须命中页面真实暴露的对象,否则返回 error、绝不执行。不许猜 selector/id/route。 - 读写分离。读路径无写工具;写工具单独成组;高危动作永远 held(默认拒绝)。
- Verify-or-refuse。每个写动作用"可观察变化"验证;
finish的 outcome(completed/failed/cancelled)由 Evidence Ledger 的证据算出,不信模型自述。 - 页面契约是地基。任何实现
data-agent-object/action/control/surface的页面,零额外代码即可被驱动。不许把某个业务页写成专用脚本作为主路径。 - 记忆只加速,不背书。Page Memory 命中就零-LLM 重放,但 verifier 始终是唯一真相;记忆失效只回退、绝不谎报或乱动;高危动作重放时仍 held。
- 协作式赌注。Attest 服务"实现契约的页面",不啃任意陌生站点。别把"无契约 DOM 降级/抓取"做成主路径(如要做,必须先 brainstorm 当独立切片)。
二、开发流程(按改动大小裁剪,别一刀切)
核心原则:流程配得上改动的大小。小改直接干,别为动几行码写一堆文档——那是负担,不是严谨。
- 小改 / 局部 / 纯函数 / 修 bug:直接做。有逻辑/边界/会回归的走 TDD(红→绿→commit);改完跑
npm test绿了就提交,不写 spec/plan。 - 只有两种情况值得先写下来:① 改核心不变量的语义(loop/verifier/ledger/memory 的判定逻辑)且方案不显然——提交前用几句话(commit body 或一小段 note)说清"为什么这么改、碰了哪条红线";② 跨多回合的大切片——才值一份简短 plan。其余一律省。
- brainstorm 只在真有分叉时用:有多条路要选、或要改架构红线的取舍时才坐下来对齐;不是每个切片的固定关卡。
- 红线不裁(§一 永远守):verify-or-refuse 不松;行为变更在声称"完成"前仍要真模型 live 验收——纯重构例外(绿的既有套件即证行为未变);诚实报告、conventional commits 照旧。
- YAGNI:赢的是做透并 ship 的一个机制,不是落不了地的宏图。能用成熟库/现成方案就用,别造轮子。
二·五、评测纪律(用户定调,2026-07-03;能力校准 2026-07-04)
- 评测只诊断、不定制。bench 的价值是暴露真实网页共性问题;任何修复必须全局性良好(对任意真实站点成立的结构理解),禁止对着某评测器的判据/词表/夹具写代码或措辞。自查题:这个修复放到一个从没见过的网站上还成立吗?不成立就不许进内核。
- 推断层新启发式必须两证齐全:合成最小例(讲清规则)+ 真实页面夹具(证明普遍性);夹具应来自多个不同站点,防单站过拟合。
- 能力校准(重要,纠正一次真实的胆怯):基础能力是入场券,不是赌注。让契约层稳稳驱动"实现了契约的标准页面"(多字段表单/下拉/多步流程/表单打开信号/handle 标签质量)是地基工作(红线4:任意实现契约的页面零额外代码可驱动),该主动、精致地投入,不算军备竞赛。要克制的只有两样:① 啃任意陌生/敌意站点(无契约 DOM 抓取当主路径);② 对着评测器凑分。除此之外,"把标准页面驱动到可靠"是分内事——别再用"能力换信任/别屠榜"把地基工作也刹住。信任没有基础能力是空的(用户定调 2026-07-04)。注意区分:提能力 ≠ 松红线;§一(verify-or-refuse/held/诚实)与本节反过拟合永不为能力让路——它们保证提升是真的、可迁移的,松了则提升变空心或说谎,自我拆台。
三、测试与验收(最重要的教训)
"全绿但崩" 是这个项目最贵的一课。 85 个 FakeLlm 测试全绿时,真模型一跑暴露出:无效写拖垮 outcome、模型省略 ref 前缀、高危 held 够不到、记忆从没记录。这些确定性测试一个都没抓到。
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.
- 8d ago First seen · 212 lines · 13,350 tokens per session scan A e260cfd129ff
Attest CLAUDE.md is an instructions file published in the GitHub repository Cyrus580529/Attest (2 stars, last pushed 2mo ago), licensed MIT. It adds 13,350 tokens to every session, about $0.0668 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 2026-08-31.
Other instructions, from other repositories
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
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).
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).
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.
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.