跳转到正文
skillsafebeta
← 报告库
用途分类 / 数据分析

Customer Research Skill 安全审计

作者说它能做什么（原文）

When the user wants to conduct, analyze, or synthesize customer research. Use when the user mentions "customer research," "ICP research," "talk to customers," "analyze transcripts," "customer interviews," "survey analysis," "support ticket analysis," "voice of customer," "VOC," "build personas," "customer personas," "jobs to be done," "JTBD," "what do customers say," "what are customers struggling

第三方安全检查结论

发现安全风险

已检查文件
4
发现的风险
3
会不会运行危险命令？检查是否下载程序后直接运行、让他人远程控制电脑，或藏起要运行的命令。未发现风险
会不会泄露文件和密钥？检查是否发送含密码或密钥的文件，以及代码里是否直接写了密钥。发现 1 项风险
中风险

客户原话和身份线索可能进入可传播的研究文档

原文依据：7 处
发现了什么

该 Skill 要求从资料和网络内容中保存精确引文、URL、日期、职位、公司规模和行业线索，并生成 VOC 引文库；没有要求删除姓名、账号、联系方式或敏感业务内容。组合这些字段可能重新识别发言者。

为什么需要注意

访谈、工单、调查或半公开社区中的客户身份、投诉和内部信息可能被复制到共享报告、营销素材或更广泛的团队文档中。

该 Skill 要求保存逐字引文、来源 URL、日期以及职位、公司规模和行业线索，并可交付 VOC 引文库。即使单个字段不敏感，这些字段组合也可能识别客户、员工或社区账号；材料中未见去标识化、敏感内容过滤、访问限制或公开内容再传播规则。风险取决于实际输入和交付对象。用户可限制保存身份线索，并要求引文脱敏及受控共享。

SKILL.md:76来自说明文档打开原文件
4. **Desired Outcomes** — what does success look like in their words?   - Capture exact quotes, not paraphrases5. **Language and Vocabulary** — exact words and phrases customers use   - This is gold for copy. "We were drowning in spreadsheets" > "manual process inefficiency"
查看另外 6 个位置
SKILL.md:144来自说明文档打开原文件
| Field | What to Capture ||-------|----------------|| Source | Platform, thread URL, date || Verbatim quote | Exact words — don't paraphrase || Context | What prompted the comment? || Sentiment | Positive / negative / neutral / frustrated || Theme tag | Pain / trigger / outcome / alternative / language || Customer profile signals | Role, company size, industry hints from the post |
SKILL.md:268来自说明文档打开原文件
1. **Research synthesis report** — themes, quotes, patterns, and implications2. **VOC quote bank** — organized verbatim quotes by theme, for use in copy3. **Persona document** — 1-3 personas built from the research4. **Jobs-to-be-done map** — functional, emotional, and social jobs by segment
references/interviews-and-surveys.md:50来自说明文档打开原文件
Recruitment methods, in order of leverage:1. **Segment the CRM** by the three signals above and pull a shortlist2. **Ask sales and CS for referrals** — they know who loves the product and who articulates why3. **Always close every call with**: *"Who else should we talk to?"* — the single most reliable way to compound your interview pipeline
SKILL.md:142来自说明文档打开原文件
For every piece of content you find:| Field | What to Capture ||-------|----------------|| Source | Platform, thread URL, date || Verbatim quote | Exact words — don't paraphrase || Context | What prompted the comment? || Sentiment | Positive / negative / neutral / frustrated || Theme tag | Pain / trigger / outcome / alternative / language || Customer profile signals | Role, company size, industry hints from the post |
SKILL.md:264来自说明文档打开原文件
## Deliverable FormatsDepending on what the user needs, offer:1. **Research synthesis report** — themes, quotes, patterns, and implications2. **VOC quote bank** — organized verbatim quotes by theme, for use in copy3. **Persona document** — 1-3 personas built from the research4. **Jobs-to-be-done map** — functional, emotional, and social jobs by segment
references/source-guides.md:340来自说明文档打开原文件
## Organizing Your ResearchUse a simple tagging system across all sources:| Tag | Meaning ||-----|---------|| `#pain` | A problem or frustration || `#trigger` | An event that prompted the search || `#outcome` | What success looks like || `#language` | Exact phrases worth using in copy || `#alternative` | Another solution they considered or use || `#objection` | Reason to hesitate or not buy || `#competitor` | Anything about a competing product |Keep a running doc with columns: Source | Date | Quote | Tags | Notes
会不会删除文件或一直在后台运行？检查是否大范围删除文件、改写磁盘，或设置自动启动。未发现风险
会不会绕过安全保护？检查是否跳过网站安全验证、开放过多文件权限，或取消操作前的确认。未发现风险
会不会误导 AI 或隐藏内容？检查工作说明是否要求 AI 忽略你的指令、干扰检查结果，或夹带看不见的文字。发现 1 项风险
中风险

要求对受访者隐瞒正式研究性质

原文依据：4 处
发现了什么

指南明确要求不要把活动称为研究或访谈，并在外联模板中表示“没有议程”。这可能使参与者无法充分理解其回答将被系统记录、分析并用于营销、产品或定位决策。

为什么需要注意

如果参与者原本会因录音、引用、商业用途或隐私原因拒绝，他们可能在缺少知情同意的情况下披露信息；这会带来信任、合规和声誉风险。

指南主动要求把研究通话描述成普通聊天，并明确禁止称其为“客户研究”；外联模板还声称“没有议程”。如果通话会被记录、分析或用于营销和产品决策，这种淡化可能使参与者在未充分理解用途的情况下提供信息。文本没有要求隐瞒录音，但也没有提供知情同意、录音告知或数据用途说明。用户可要求作者加入明确的研究目的、记录方式、用途和退出机制。

references/interviews-and-surveys.md:14来自说明文档打开原文件
> The first rule of customer research: you do not talk about customer research.Keep it casual. The moment a customer thinks they're in "a research study" they perform — they give you the polished, socially-acceptable answer instead of the real one. Frame calls as a chat, not an interview. Don't lead. Don't pitch. Don't defend the product. You're there to listen and learn how they actually think, talk, and decide.
查看另外 3 个位置
references/interviews-and-surveys.md:63来自说明文档打开原文件
Keep it short, casual, specific, and low-commitment. Not a "research study."
references/interviews-and-surveys.md:70来自说明文档打开原文件
I'm [name] from [company]. I'm trying to get better at helping customerslike you, and I'd love to steal 30 minutes to hear how [product area] isactually working for you — what's good, what's annoying, what you wishwere different. No pitch, no agenda.As a thank you I'll send you a $50 [Amazon/Visa] gift card.
references/interviews-and-surveys.md:83来自说明文档打开原文件
Notes:- "No pitch, no agenda" and "what's annoying" signal you actually want the truth.- One clear ask, two concrete time options, a booking link. Remove friction.- Never say "customer research study."
会不会偷偷改推广链接或收款方？检查是否强制替换推广链接或收款对象，同时要求隐瞒更改。发现 1 项风险
中风险

用“最佳客户”样本评估 PMF 可能系统性高估市场契合度

原文依据：4 处
发现了什么

访谈指南先要求优先选择高客单价、销售周期短且低流失的最佳客户，随后把单一调查中“非常失望”达到 40% 视为可能已有 PMF，并把该群体称为“真正的市场”。如果同一偏好样本用于 PMF 调查，结果会天然偏高。

为什么需要注意

用户可能基于失真的 PMF 指标扩大投入、调整路线图或忽略流失及普通用户的问题，造成错误的产品和预算决策。

这段证据能说明什么

指南确实要求访谈优先选择高客单价、短销售周期、低流失的客户，同时给出 40% PMF 阈值并称“非常失望”群体为真正市场。若把同一偏好样本用于 PMF 调查，结果可能偏高并影响路线图决策。然而原文没有明确要求 PMF 调查只发给这批“最佳客户”，且承认访谈有选择偏差，因此无法确认该采样错误是规定行为。用户应要求作者明确 PMF 调查的总体、抽样方式和分群报告。

这项判断针对展示的代码和适用条件，不表示风险已经实际发生。
references/interviews-and-surveys.md:42来自说明文档打开原文件
### Recruit your best customersDon't interview whoever answers first. Interview the customers you want *more of*. Segment your CRM and prioritize by:- **High deal size** — the accounts worth the most- **Short sales cycle** — they "got it" fast; their language converts fast- **Low churn / high retention** — they got real, lasting valueRecruitment methods, in order of leverage:1. **Segment the CRM** by the three signals above and pull a shortlist2. **Ask sales and CS for referrals** — they know who loves the product and who articulates why3. **Always close every call with**: *"Who else should we talk to?"* — the single most reliable way to compound your interview pipeline
查看另外 3 个位置
references/interviews-and-surveys.md:125来自说明文档打开原文件
**The 40% benchmark**: if **40% or more** of users answer **"very disappointed,"** you likely have product/market fit. Below 40%, keep iterating. **Superhuman reached 58%** by engineering their roadmap around this metric — segmenting on the "very disappointed" cohort, doubling down on what that cohort loved, and converting the "somewhat disappointed" fence-sitters.Run it as a recurring pulse, not once. Follow the core question with:- *"What type of person do you think would most benefit from [product]?"* (sharpens ICP)- *"What is the main benefit you receive from [product]?"* (your positioning, in their words)- *"How can we improve [product] for you?"* (roadmap fuel from fence-sitters)Segment every answer by the "very disappointed" cohort vs. the rest — that cohort is your true market.
references/interviews-and-surveys.md:127来自说明文档打开原文件
Run it as a recurring pulse, not once. Follow the core question with:- *"What type of person do you think would most benefit from [product]?"* (sharpens ICP)- *"What is the main benefit you receive from [product]?"* (your positioning, in their words)- *"How can we improve [product] for you?"* (roadmap fuel from fence-sitters)Segment every answer by the "very disappointed" cohort vs. the rest — that cohort is your true market.
references/source-guides.md:366来自说明文档打开原文件
| Source | Signal Strength | Bias to Note ||--------|----------------|--------------|| Customer interviews (unprompted) | Very high | Small sample; selection bias toward engaged customers || Win/loss interviews | High | Recent memory only; rationalization common || App store / G2 reviews | High | Skews toward strong opinions (love or hate) |

Skill 逻辑拆解

8 个说明模块

该 Skill 会先自动读取项目中的产品营销上下文文件，再决定需要询问用户哪些信息；这些文件可能包含内部定位、客户或商业资料，但所示文本没有要求把内容发送到外部。

查看原文
SKILL.md:14来自说明文档打开原文件
**Check for product marketing context first:**If `.agents/product-marketing.md` exists (or `.claude/product-marketing.md`, or the legacy `product-marketing-context.md` filename, in older setups), read it before asking questions. Use that context to skip questions already answered.

它支持分析用户提供的访谈、调查、评论和工单，也支持从 Reddit、G2、论坛及社区收集公开信号，并记录原话、来源、日期、情绪和人物特征。

查看原文
SKILL.md:21来自说明文档打开原文件
### Mode 1: Analyze Existing AssetsYou have raw research material (transcripts, surveys, reviews, tickets). Your job is to extract signal.### Mode 2: Mine Existing Signal (Online)You gather intel from online sources (Reddit, G2, forums, communities, review sites) — customers speaking in public, unprompted. Your job is to know where to look and what to extract.
SKILL.md:144来自说明文档打开原文件
| Field | What to Capture ||-------|----------------|| Source | Platform, thread URL, date || Verbatim quote | Exact words — don't paraphrase || Context | What prompted the comment? || Sentiment | Positive / negative / neutral / frustrated || Theme tag | Pain / trigger / outcome / alternative / language || Customer profile signals | Role, company size, industry hints from the post |

该 Skill 没有提供安装命令或可执行脚本；其主要输出是研究主题、引文、人物画像、待办任务和竞争情报等文档，并要求在生成前询问用户所需交付物。

查看原文
SKILL.md:264来自说明文档打开原文件
## Deliverable FormatsDepending on what the user needs, offer:1. **Research synthesis report** — themes, quotes, patterns, and implications2. **VOC quote bank** — organized verbatim quotes by theme, for use in copy3. **Persona document** — 1-3 personas built from the research4. **Jobs-to-be-done map** — functional, emotional, and social jobs by segment5. **Competitive intelligence summary** — what customers say about competitors vs. you6. **Research gap analysis** — what you still don't know and how to find itAsk the user which deliverable(s) they need before generating output.
从这里开始 · 工作说明SKILL.md
customer-research
连线表示工作说明包含的模块，不是实际运行顺序。点击模块可查看原文。 另有 1 个章节，可在原文件中查看。

文件引用关系图

4 处引用
哪些文件发起引用引用了什么
连线表示真实的文件引用，不是运行顺序。点击节点可高亮相关连线，并查看具体文件和原文位置。虚线表示还有文件需要定位。
文件与检查记录4 个文件

检查范围与遗漏

逐文件查看涉及的内容

下方列出本次涉及的原文范围；纳入检查不代表已查清所有问题。

  • SKILL.md已纳入全文
  • references/interviews-and-surveys.md已纳入全文
  • references/source-guides.md已纳入全文
  • evals/evals.json已纳入全文

这份报告只针对上方版本。我们看了拿到的代码和说明文件，没有实际运行 Skill，也没有检查它另外安装的软件包。因此，这不是“保证安全”的承诺；换了版本或使用环境，结果也可能不同。

  • SKILL.md工作说明
  • evals/evals.json配套文件
  • references/interviews-and-surveys.md配套文件
  • references/source-guides.md配套文件

代码和说明中提到的操作

读取文件
references/interviews-and-surveys.md:6来自说明文档打开原文件
Three primary-research pillars, best used together:1. **Video calls** — deep, unstructured, follow-the-thread (this file)2. **Surveys** — broad, quantified, benchmarkable (this file)
连接外部网站
references/source-guides.md:15来自说明文档打开原文件
- Search `site:reddit.com "[job title] tools"` or `site:reddit.com "[problem category] software"`- Use [subreddit search tools](https://www.reddit.com/subreddits/search) with problem-space keywords- Look at what subreddits show up in Google results when you search ICP problems
读取了多少行
1,041
文件校验值（用于核对版本）
94d6f699c6ea247a0b9602b099b64ba4567ba48dd15654cf19bdc81d780d1bd8