菜单
Plans a 5-person user interview sprint and synthesizes findings. Use before building or whenever you are guessing about users instead of knowing. Produces research-report.md.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
Explains your product's technical architecture in plain English — so you understand what's being built, can make informed decisions, and can describe it to others. Use when an agent has set up your tech stack and you're not sure how the pieces connect, or when you need to explain your system to a potential co-founder, investor, or hire. Produces an architecture-overview.md in plain language with a simple diagram.
Prioritizes what to build in what order, given your MVP brief and tech stack. Use when you have a feature list and don't know where to start, or when you're building in the wrong order and feeling stuck. Produces a build-sequence.md with a sequenced backlog and a clear "build this first" decision.
Recommends the specific third-party services to use for auth, payments, email, storage, and other common needs. Use when building a feature that requires an external service and you're unsure which tool to use or how to set it up. Produces an integrations-plan.md with specific tool choices and the order to add them.
Recommends the right technology stack for your product, branching by product type (consumer mobile, B2B SaaS, internal tool, API/developer tool, marketplace, AI-native). Use before starting to build, when you're unsure what tools to use, or when an agent has suggested a stack that feels overwhelming. Produces a stack-decision.md with specific tool recommendations and setup order.
The recurring founder ritual — run after every meaningful ship or weekly, whichever comes first. Guides you through honest reflection, user signal synthesis, MPP/PMF assessment, and one clear decision for the next cycle. Produces a dated cycle record that compounds over time. Use whenever you've shipped something and want to figure out what it taught you and what to do next.
Diagnoses why your product isn't gaining traction after multiple build cycles. Spawns 5 parallel hypothesis agents — one per failure mode — each gathering independent evidence. Synthesizes into a ranked diagnosis with a specific 2-cycle prescription. Triggered automatically by build-cycle after 3+ flat cycles, or invoked directly when stuck. Produces diagnosis.md.
基于 SOC 职业分类
| name | user-interview-guide |
| description | Plans a 5-person user interview sprint and synthesizes findings. Use before building or whenever you are guessing about users instead of knowing. Produces research-report.md. |
| phase | strategy |
| version | 1.0.0 |
Default: answer in chat first with Verdict, Do this now, and Details saved.
Keep chat under 150 words unless asked; do not paste the full artifact.
Write/update the requested .md file as the durable record for agents.
Say: "Help me interview users" or "I need to talk to real users before I build this"
Share your product idea, customer profile, or the assumption you want to test.
Output: research-report.md — themes, surprises, and verbatim quotes synthesized from 5 user interviews.
A research-report.md containing: the top 3 themes across all interviews, the top 3 surprises that challenge your assumptions, and the 5 best verbatim quotes that capture the insights. Plus the interview plan, recruiting message, and facilitation guide you used to get there.
Example output excerpt: Theme #1: Freelance designers don't track time — they track energy. 4 of 5 interviewees described their workflow in terms of "deep focus blocks" rather than hours. None used a timer. Two had tried Toggl and abandoned it within a week. Best quote: "I don't bill by the hour. I bill by the deliverable. So tracking time feels like I'm lying to myself about how I work." — P3, freelance brand designer, 6 years experience.
This skill embeds the principles of The Mom Test by Rob Fitzpatrick. Every question, every guideline, and every warning in this skill is designed to produce good data and avoid bad data.
Good data:
Bad data:
Core rules:
Generate the interview plan. Auto-read customer-profile.md if it exists in the project — it shapes who to recruit and what to ask.
These are starter questions. Adapt the language to fit your domain, but preserve the structure — every question asks about the past, asks for specifics, and never reveals your idea.
"Walk me through the last time you [did the thing your product addresses]. What happened?" Opens with a concrete, recent event. Not hypothetical.
"What was the hardest part of that?" Surfaces the real pain, in their words.
"Why was that hard?" Goes one level deeper. The first answer is rarely the real answer.
"How are you dealing with that problem today?" Reveals current workarounds. If they have none, the pain may not be real.
"What have you tried before to solve this? What happened?" Shows whether they have spent time or money searching for a solution.
"What don't you love about the tools/process you're using now?" Surfaces gaps in existing solutions without leading toward yours.
"If you could wave a magic wand and fix one thing about [this process], what would it be?" Use sparingly — this can produce hypotheticals. Follow up immediately with: "Tell me about the last time that thing caused a real problem."
"Is there anything else about [this area] that I should have asked about but didn't?" Often produces the most valuable insight of the entire interview.
Subject: Quick chat about how you [do the relevant activity]?
Hi [Name],
I'm researching how [target role/persona] handle [the activity or problem area].
I'm not selling anything — I'm just trying to understand the workflow better
before I build something wrong.
Would you be open to a 20-minute call this week? Happy to work around your
schedule. I'd love to hear how you actually do [the thing] day-to-day.
[Your name]
If customer-profile.md exists, use it to define the recruiting criteria. If not, define:
Where to find them: LinkedIn, relevant Slack/Discord communities, Twitter/X, Reddit, professional associations, personal network (second-degree connections preferred over friends).
Step-by-step facilitation guide for conducting each interview.
Warm-Up (3 minutes)
Core Questions (15 minutes)
Wrap-Up (2–5 minutes)
| Signal | What It Means | Write It Down |
|---|---|---|
| Specific story with details | Real pain, real experience | Verbatim quote + context |
| Emotional language (frustrated, embarrassing, hate) | Strong pain signal | The exact words they used |
| Money or time already spent | Validated willingness to pay/invest | Amount + what they tried |
| Workaround they built | Problem is real enough to act on | What they built and why |
| "It's not that big a deal" | Weak pain — may not be worth solving | Note it honestly |
| Compliments about your idea | Bad data — they are being polite | Discard. Do not count this. |
After the founder completes 5 interviews and pastes raw notes, spawn 3 agents simultaneously to analyze the data.
The founder pastes raw interview notes (all 5 interviews). The orchestrator passes the complete set of notes to all 3 agents.
Agent A — Pattern Finder Context: All 5 interview transcripts/notes Task: Identify the top 3 themes that appear across multiple interviews. For each theme:
Returns: Top 3 themes, ranked by frequency and strength of evidence.
Agent B — Surprise Extractor
Context: All 5 interview transcripts/notes + customer-profile.md or founder's stated assumptions (if available)
Task: Identify the top 3 unexpected findings — things that challenge the founder's assumptions or reveal something the founder did not anticipate.
For each surprise:
Returns: Top 3 surprises, ranked by potential impact on product decisions.
Agent C — Quote Curator Context: All 5 interview transcripts/notes Task: Select the 5 best verbatim quotes from across all interviews. Selection criteria:
For each quote:
Returns: 5 curated quotes with attribution and significance.
Wait for all 3 agents. The orchestrator synthesizes.
# Research Report — [YYYY-MM-DD]
## Interview Summary
- **Interviews completed:** 5
- **Participant profiles:** [Brief description of each: P1–P5]
- **Date range:** [When interviews were conducted]
## Top 3 Themes
### Theme 1: [Theme Name]
**Frequency:** [X] of 5 interviewees
**Evidence:** [Agent A's strongest evidence]
**Implication:** [What this means for the product]
### Theme 2: [Theme Name]
[Same format]
### Theme 3: [Theme Name]
[Same format]
## Top 3 Surprises
### Surprise 1: [What was unexpected]
**Assumed:** [What the founder believed]
**Reality:** [What interviews revealed]
**Source:** [Which participants, P1/P3/P5]
**Impact:** [How this changes the product direction]
### Surprise 2: [What was unexpected]
[Same format]
### Surprise 3: [What was unexpected]
[Same format]
## 5 Key Quotes
> "[Verbatim quote]"
> — P[X], [role], [context]
> **Why it matters:** [One sentence]
> "[Verbatim quote]"
> — P[X], [role], [context]
> **Why it matters:** [One sentence]
[...3 more quotes...]
## Recommended Next Steps
1. [Action based on Theme 1]
2. [Action based on Surprise 1]
3. [What to test or validate next based on findings]
## Raw Interview Notes
[Appended or linked — preserve the original data]
If your agent does not support parallel subagents, run the synthesis in sequence:
Then assemble research-report.md.
Same output. ~3x longer.
Founder: Building a project scoping tool for freelance designers who struggle to estimate timelines and price projects accurately.
Phase 1 — PLAN:
customer-profile.md exists: "Freelance UI/UX designers, 2–7 years experience, solo or small studio, earning $60K–$150K/year, primarily finding clients through referrals and Dribbble."Phase 2 — GUIDE:
Phase 3 — SYNTHESIZE:
Results:
Theme #1: Designers undercharge because they cannot articulate scope, not because they do not know their value. 5 of 5 interviewees described situations where they knew the project was worth more but could not break it down convincingly for the client. P1: "I know it's a $10K project but I can't explain why, so I say $6K and hate myself."
Surprise #1: Nobody wants a scoping "tool" — they want a script. The founder assumed designers needed software to calculate estimates. But 4 of 5 said their real problem was the client conversation itself — they freeze when the client pushes back on price. They wanted words, not math.
Quote: "I don't need a calculator. I need someone to tell me what to say when the client goes 'that seems expensive.'" — P4, freelance product designer, 4 years experience.
The founder pivoted from building a scoping calculator to building a guided proposal conversation tool with response scripts for common client objections.