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 skills add Lambenthan/paper-discipline-skills --skill paper-confirm-before-doinggit clone --depth 1 https://github.com/Lambenthan/paper-discipline-skillsWrote 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/skills/lambenthan/paper-discipline-skills/paper-confirm-before-doing)<a href="https://agentmods.dev/skills/lambenthan/paper-discipline-skills/paper-confirm-before-doing"><img src="https://agentmods.dev/badge/skills/lambenthan/paper-discipline-skills/paper-confirm-before-doing/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/skills/lambenthan/paper-discipline-skills/paper-confirm-before-doing"><img src="https://agentmods.dev/badge/skills/lambenthan/paper-discipline-skills/paper-confirm-before-doing.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 | $0.00146 | $0.01418 |
| Opus 5 | $0.00073 | $0.00709 |
| Sonnet 5 | $0.00029 | $0.00284 |
| Haiku 4.5 | $0.00015 | $0.00142 |
Grade A, and why
paper-confirm-before-doing 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 12d 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.
What it actually says
paper-confirm-before-doing：先确认方案，再动手
核心理念
模糊任务最贵的不是执行成本，是做完发现不对再撤销的代价。 口头一句确认 = 30 秒；写完发现跑偏要回滚 = 半小时起。
你的 AI 越能干，越要在动手前停下来对一次表。
触发条件（命中任一即触发）
用户的请求包含以下模式：
- 宽泛动词：改、整理、润色、重写、优化、清理、统一、规范化、处理一下、看一下、帮我搞下
- 范围未指定：没说改哪几节、改哪几段、改哪几个文件
- 方法未指定：没说按什么标准改、保留什么、删掉什么
- 目标读者未指定：没说给导师看 / 给评审看 / 给毕业答辩用
- 量词模糊：「全部」「所有」「都」却没明确指哪个集合
只要命中任意一条，立刻进入"先确认"模式。 不要因为任务看起来小就跳过——书里的翻车案例 80% 是从"小活儿"开始的。
强制流程
用户给出模糊任务
│
▼
立即停下，禁止调用任何 Edit / Write / Bash 写操作
│
▼
在回答里写出四件事：
1. 我理解的任务边界（改哪里、不改哪里）
2. 我打算用的方法（按什么标准、保留与删除原则）
3. 一个最小样本（先改一段示例 / 一节 / 列一个示意）
4. 显式问句：「方案行的话，我就推全文。可以吗？」
│
▼
明确等待用户回复
│
▼
用户确认 → 按方案执行
用户调整 → 改方案后重新确认（不是改方案后直接做）
标准回复模板
我理解的任务是：[具体范围]。 准备这样做：[方法、标准、保留与删除原则]。 不会动的部分：[显式列出]。 先给你一个小样本：[一段 / 一节 / 一处示意]。 你看方案行的话，我就按这个推全文。可以吗？
❌ 反例（书 §3.2 真实案例）
用户：「帮我把第三章润色一下。」
错误做法：直接调用 Edit 改了 1500 字，结果发现：
- 把作者特意保留的口语化表达改没了
- 把导师让保留的引文改写了
- 改完用户要求恢复，但中间结果没存
正确做法：先回复
我准备按学术正式度统一术语、压缩冗余、保留你原文的段落结构和小标题。 不会动的：图表说明、引文原文、术语缩写表。 先改 §3.2.1 第一段给你看效果，确认后推全章——可以吗？
Rationalization Table
| 念头 | 现实 |
|---|---|
| "用户语气很急，先做了再说" | 急的反义词不是快做，是少返工 |
| "这种小活儿不用确认" | 没有不用确认的活儿，只有还没翻车的活儿 |
| "用户上次也是这种任务，我知道他要啥" | 上次不是这次，章节不是同一章 |
| "我先做一遍，他不满意再改" | "再改"在 AI 时代是从头来过，不是局部调整 |
| "确认太啰嗦了，影响体验" | 改错了再撤销才影响体验 |
| "我理解很到位，不会跑偏" | 你理解到位 = 你能用一句话说清——那就说出来等他点头 |
| "用户已经发了三次类似任务了" | 三次类似 ≠ 第四次相同 |
Red Flags：以下情况 = 立即停止，回去确认
- 你已经在脑子里规划了改哪几个文件 → 停，先告诉用户
- 你正要打开 Edit / Write 工具 → 停，先告诉用户
- 你想"先改一点试试" → 停，先把"试试"的方案告诉用户
- 你觉得"这显然就该这么改" → 显然 = 用户和你想的一样 = 必须确认
- 你打算"边改边问" → 不行。先全部说完，再等回复
例外（极少数情况可以跳过确认）
只有以下三种情形可以直接动手：
- 用户明确写出了完整方案："把这五段里的『论证』改成『论述』，其他不动。"
- 纯查询类：用户问"这段在哪个文件里"、"这个引用对不对"——不修改任何文件
- 回答用户的提问：不涉及写文件 / 跑批量
任何不属于以上三种的修改请求 = 必须确认。
来源
《Claude Code 科研手记》§3.2「先确认方案，再动手」
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.
- 12d ago First seen · 127 lines · 146 tokens per session scan A 7b2ac60e38f4
paper-confirm-before-doing is a skill published in the GitHub repository Lambenthan/paper-discipline-skills (19 stars, last pushed 4mo ago), licensed MIT. It adds 146 tokens to every session and 1,418 once invoked, about $0.0007 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 skills, from other repositories
secure-auth
Secure authentication patterns (OWASP, NIST). Use for login, registration, password reset, sessions, JWT, OAuth, MFA, passkeys.
academic-writing
Scholarly writing and research compliance. Use for CRediT, preregistration, Plan S, Nelson Memo, preprints, ORCID, LLM disclosure.
page-monitoring
Web page change detection, availability tracking, and RSS feed generation. Use to monitor changes, downtime, or make a feed.
web-archiving
Web archiving and retrieval via Wayback Machine and Archive.today. Use to preserve content, reach dead pages, or save evidence.
accessibility-compliance
Web accessibility patterns for news and academic sites. Use for WCAG audits, alt text, accessible data viz, and assistive tech.
one-way-door
Flags irreversible decisions before commit. Use for data models, infra, auth boundaries, API contracts, event schemas, CI/CD.