customer-discovery-conversations
Plan and audit customer discovery conversations that produce concrete evidence instead of compliments, opinions, hypotheticals, or feature-request noise. Use when planning customer interviews, rewriting interview scripts, validating startup ideas before building, avoiding biased questions, or applying Mom Test-style customer learning.
- License 许可
- MIT
- 最近更新
- 2026-11-14
- Stars
- ★ 1.6k
- Forks
- 491
01 前言 Front Matter
- name
- customer-discovery-conversations
- license
- MIT
- metadata
- { "tags": "customer-discovery,customer-interviews,startups,validation,research", "version": "0.1.3", "category": "Marketing", "displayName": "Customer Discovery Conversations" }
- description
- Plan and audit customer discovery conversations that produce concrete evidence instead of compliments, opinions, hypotheticals, or feature-request noise. Use when planning customer interviews, rewriting interview scripts, validating startup ideas before building, avoiding biased questions, or applying Mom Test-style customer learning.
- compatibility
- Codex, Claude Code, and other Agent Skills-compatible clients.
02 资源文件 4
- 参考文档agents/openai.yaml208 B
- 参考文档workflows/recover-bad-data.md1010 B
- 参考文档workflows/rewrite-interview-questions.md1.5 KB
- 参考文档workflows/run-conversation-batch.md1.3 KB
03 SKILL.md 正文
Customer Discovery Conversations
Use this skill to turn customer conversations into decision-quality evidence. The goal is not approval for an idea. The goal is to learn concrete facts about customers' lives, current behavior, constraints, costs, workarounds, and risks.
Source Traceability
Primary source: The Mom Test by Rob Fitzpatrick, especially chapters 1-4 and 8. Guidance is paraphrased for this MIT repo; authoring notes used converted EPUB lines 236-2362 and 3613-4108.
Workflow Routing
| Need | Use |
|---|---|
| Rewrite an interview script | workflows/rewrite-interview-questions.md |
| Run a batch of conversations | workflows/run-conversation-batch.md |
| Recover from compliments, hypotheticals, or ideas | workflows/recover-bad-data.md |
| Segment is too broad | customer-segment-slicing |
| Product or sales meeting needs a real next step | customer-commitment-validation |
| Raw notes need synthesis | customer-learning-notes |
Conversation Workflow
- State the decision the conversations should improve.
- Name the riskiest assumptions, including at least one uncomfortable question.
- Choose the narrow customer or stakeholder type. If the segment is fuzzy, use
customer-segment-slicingfirst. - Prepare up to three learning goals for this type of person.
- Rewrite questions toward current behavior and specific past examples.
- Keep the conversation about the customer until there is enough evidence to discuss a solution.
- During the conversation, anchor vague claims to recent examples and dig into current costs, workarounds, constraints, and goals.
- Capture notes, signals, and follow-up tasks for review.
Question Rules
- Talk about the customer's life, work, tools, constraints, and recent past.
- Ask scary questions early enough that the answer can still change the plan.
- Turn opinions, compliments, and hypotheticals into requests for examples.
- Ask what already costs them time, money, reputation, or operational pain.
- Save the pitch until the customer evidence justifies discussing a solution.
Good Evidence
Prefer recent examples, current workarounds, money/time/reputation already spent, existing tools and budgets, named stakeholders, decision paths, and strong emotion backed by a concrete story.
Treat compliments, opinions, future promises, generic claims, and feature requests without motivation as weak evidence.
Output Format
# Customer Discovery Plan
## Context
- Decision:
- Segment:
- Risk questions:
## Conversation Guide
| Goal | Ask About | Avoid | Follow-Up |
|------|-----------|-------|-----------|
## Evidence To Capture
- Current behavior, past examples, workarounds:
- Costs, budget, constraints, stakeholders:
- Strong emotion, follow-up, commitment:
## Review Criteria
- What changed in our beliefs?
- What is still unknown?
- What are the next three questions?
Quality Bar
- Do not help users collect approval or compliments.
- Do not ask customers to design the product for the team.
- Do not treat "I would buy this" as validation without commitment.
- Do not run conversations without a decision, segment, and learning goals.
- Prefer fewer high-value conversations over many generic interviews.
04 同仓库更多 Skills
worktable-review-with-annotations
Create, inspect, reply to, update, or resolve situated Worktable annotations. Use when feedback belongs to a Doc section, block, text range, or HTML Doc, or when the user asks to address existing review feedback and coordinate a requested artifact edit.
worktable-find-and-synthesize-context
Find, assess, and synthesize context from Worktable Docs and Records. Use when the user asks what a workspace contains, wants workspace-backed research, needs current facts reconciled across artifacts, or requests a cited synthesis without changing workspace state.
worktable-create-or-update-docs
Create, edit, rename, or maintain narrative Worktable Docs. Use for plans, decisions, briefs, notes, research syntheses, and other durable prose when the user requests a new artifact or a safe update to an existing one.
worktable-create-or-manage-records
Organize repeated structured information as durable Worktable Records. Use for research datasets, inventories, directories, pipelines, trackers, and other collections whose items need identity, fields, filtering, links, validation, or individual updates.
worktable-create-or-manage-interactive-html
Create or improve durable visual and interactive artifacts in Worktable. Use for dashboards, trackers, calculators, mini apps, slide presentations, and other requests that benefit from a polished HTML Doc.
worktable-collaborate-threads
Read, answer, and hand off messages through durable Worktable threads with users or other authenticated participants. Use when the user asks to check addressed messages, reply to a participant, continue a thread, request another agent's help, wait briefly for a response, or watch for comments and messages sent to the agent.