Token导航 LogoToken导航

Agent Skill · 研究检索

foundation-lean-canvas(基础精益画布)

foundation-lean-canvas 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

foundation-lean-canvas 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为操作浏览器;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 67,870。

总安装 240周安装 10GitHub Stars 168下载 80核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:foundation-lean-canvas(基础精益画布)
来源仓库:https://github.com/product-on-purpose/pm-skills
仓库路径:skills/foundation-lean-canvas
安装命令:
npx skills add https://github.com/product-on-purpose/pm-skills --skill foundation-lean-canvas
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/product-on-purpose/pm-skills --skill foundation-lean-canvas

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

foundation-lean-canvas 用于查找、检索和筛选相关信息。适合在 Codex、Claude、Cursor、Gemini CLI 中根据任务场景定位内容。通过 npx skills add 命令从指定仓库安装并使用该技能。安装前需确认权限范围、维护状态及是否触发联网或文件操作。

建议结合原始 README 核验具体用法和功能边界。

来源仓库内路径:skills/foundation-lean-canvas

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/product-on-purpose/pm-skills --skill foundatio…;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要操作浏览器,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

foundation-lean-canvas 怎么安装?

在终端执行 npx skills add https://github.com/product-on-purpose/pm-skills --skill foundation-lean-can…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

foundation-lean-canvas 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

foundation-lean-canvas 需要什么权限?

权限级别是操作浏览器。该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

foundation-lean-canvas 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

foundation-lean-canvas 用的人多吗?

公开数据是累计安装 240 次,近一周安装 10 次,下载 80 次,GitHub 168 星。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

foundation-lean-canvas 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
foundation-lean-canvas

基础精益画布 · 本页

240168操作浏览器未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Lean Canvas

A lean canvas is a one-page business thesis that makes your assumptions about problem, customer, solution, and viability explicit and testable. Developed by Ash Maurya from Alex Osterwalder's Business Model Canvas, it is specifically adapted for startups and product teams operating under uncertainty. Nine interlocking blocks force you to articulate the whole picture at once so that changing one block's assumptions surfaces the ripple effect on the others.

This skill is a strategic hub, not a specialist tool. It produces the integrated one-page artifact and cross-links to deeper PM skills (/problem-statement, /persona, /jtbd-canvas, /solution-brief, /competitive-analysis, /experiment-design) for single-block depth when needed.

Supported Modes

  • content (default) produces the nine-block canvas as structured markdown.
  • visual produces the markdown canvas AND writes a self-contained, attractive .html file to disk using references/html-template.html as the layout scaffold. The HTML renders the canonical Maurya nine-block layout with polished typography, subtle per-column color accents, confidence badges per block, and print-ready A3 landscape styling. No external assets or CDN dependencies: the file opens correctly in a browser with no network access.

If mode is omitted, default to content and state that fallback explicitly.

When to Use

  • Framing a new product, feature, or business thesis on one page
  • Stress-testing an existing business by making implicit assumptions explicit
  • Comparing two or more strategic options side-by-side (run the skill once per option, then diff)
  • Onboarding new team members into the strategic thesis in a single artifact
  • Mid-phase reality check: does the thesis still hold given what we have learned?
  • Pairing with /experiment-design to prioritize which block assumptions to test first

When NOT to Use

  • You need deep research on a single block (persona detail, problem framing, competitive landscape). Use the specialist skill (/persona, /problem-statement, /competitive-analysis) instead.
  • You are drafting a PRD, user stories, or acceptance criteria. Use /prd, /user-stories, /acceptance-criteria; lean canvas is strategy, not specification.
  • You want to brainstorm solutions without a customer-problem anchor. Start with /problem-statement or /jtbd-canvas and return to lean canvas once the problem is framed.
  • You need a Business Model Canvas for an established enterprise with known customers and channels. Maurya designed lean canvas specifically for high-uncertainty early-stage ventures; a BMC is a better fit for steady-state analysis.

Instructions

When asked to create a lean canvas, follow these steps:

  1. Resolve mode and intent Determine whether the request is content or visual. If mode is omitted, default to content and state the fallback. Clarify the target: new product thesis, existing-business stress test, or side-by-side comparison of options. If unclear, ask once before proceeding.
  2. Collect context and evidence Use user-provided context first: product name, market, target customer, any research already done, existing alternatives users are hiring today, known constraints. If evidence is thin, continue generation but mark gaps in the Evidence & Confidence section and calibrate per-block confidence accordingly. For existing businesses, distinguish current assumptions from validated data explicitly.
  3. Fill the nine blocks in recommended order Fill in this order because each block's answer constrains the next. Do not skip ahead. a. Problem. Top 3 problems, ranked by pain intensity and frequency. Include Existing Alternatives (what customers do today, including workarounds and non-consumption, not just direct competitors). b. Customer Segments. Who has these problems most acutely? Name Early Adopters as a distinct subset you will reach first. Early Adopters are more painful, more reachable, and more willing to try a new solution than the broader segment. c. Unique Value Proposition (UVP). One sentence that makes a clear, testable promise. Include a High-Level Concept ("X for Y" analogy) that accelerates understanding for busy readers. d. Solution. Top 3 features that address the top 3 problems. Map 1:1 to the Problem block. Keep it concrete but do not over-engineer; this is a hypothesis, not a spec. e. Channels. Free and paid paths to your early adopters. Distinguish compounding channels (content, SEO, community) from traction-demonstrating channels (outbound, paid ads). f. Revenue Streams. Model (subscription, transaction, freemium, services), price point, expected volume, and LTV. Show the math so the revenue thesis is inspectable. g. Cost Structure. CAC, fixed vs variable, and the cost driver that shapes the growth curve. h. Key Metrics. The 3 to 5 leading indicators that signal whether the model is working. AARRR (Acquisition, Activation, Retention, Revenue, Referral) is a useful default frame. i. Unfair Advantage. What cannot be easily copied or bought. Empty is acceptable if framed as an open question; never fabricate a moat.
  4. Apply evidence and confidence policy Tag each block with High, Medium, or Low confidence plus a one-line rationale. Populate the Evidence & Confidence section: Validated (assumptions with named sources), Assumed (no data yet), Open Questions (what you would need to learn to raise confidence), Governance (who owns the canvas and when it is revisited). A block marked "High" must name a specific evidence source, not a generic claim.
  5. Render and write the visual file (visual mode only) Read references/html-template.html. It is a complete, self-contained HTML5 document using CSS Grid to express the canonical Maurya nine-block layout, with per-column color accents, confidence badges, and A3 landscape print styling. Fill every {{PLACEHOLDER}} token in the template with content from the markdown canvas:

- {{PRODUCT_NAME}}, {{CREATED_DATE}}, {{PURPOSE}}, {{OVERALL_CONFIDENCE}} from the canvas header. - {{PROBLEM_CONTENT}}, {{EXISTING_ALTERNATIVES}}, {{SOLUTION_CONTENT}}, {{UVP_CONTENT}}, {{HIGH_LEVEL_CONCEPT}}, {{ADVANTAGE_CONTENT}}, {{CUSTOMER_CONTENT}}, {{EARLY_ADOPTERS}}, {{METRICS_CONTENT}}, {{CHANNELS_CONTENT}}, {{COST_CONTENT}}, {{REVENUE_CONTENT}} from the respective blocks. Use a <ul><li> list for multi-item blocks; keep cell content concise (one-line summaries, not the full markdown detail) so the visual stays scannable. - {{CONF_PROBLEM}}, {{CONF_SOLUTION}}, {{CONF_UVP}}, {{CONF_ADVANTAGE}}, {{CONF_CUSTOMER}}, {{CONF_METRICS}}, {{CONF_CHANNELS}}, {{CONF_COST}}, {{CONF_REVENUE}} with the single letter H, M, or L matching each block's confidence tag. Each appears twice in the template (once in the class attribute for styling, once as visible text). - {{VALIDATED_COUNT}}, {{ASSUMED_COUNT}}, {{OPEN_QUESTIONS_COUNT}}, {{OWNER}}, {{NEXT_REVIEW}} in the footer evidence strip. Write the filled document to disk at a user-specified path, or default to ./lean-canvas-{slug}.html in the current working directory (where {slug} is the product name lowercased with non-alphanumeric characters replaced by hyphens). Do not introduce external font or CSS links; the template is intentionally self-contained. In content mode, skip this step entirely.

  1. Finalize for direct use Remove all template guidance blockquotes (> notes) from the final artifact. Verify UVP is decision-changing and testable, not marketing fluff. Confirm Early Adopters is a distinct subset (not a restatement of Customer Segments). Confirm Solution block items map 1:1 to Problem block items.

Output Contract (v1.0.0)

  • All nine blocks present in the canonical order (Problem, Customer Segments, UVP, Solution, Channels, Revenue Streams, Cost Structure, Key Metrics, Unfair Advantage)
  • Each block contains content, a confidence tag (High|Medium|Low), and a one-line rationale
  • Evidence & Confidence section present with Validated, Assumed, Open Questions, and Governance subsections populated (even if a subsection is intentionally empty, mark it "None" rather than removing it)
  • Content mode: the markdown canvas is the single output.
  • Visual mode: in addition to the markdown canvas, a self-contained HTML file is written to disk at ./lean-canvas-{slug}.html (or a user-specified path). The HTML is derived from references/html-template.html with every placeholder filled and no external dependencies. After writing, report the file path to the user.
  • Foundation classification: no phase: field in frontmatter; uses classification: foundation

Quality Checklist

Before finalizing, verify:

  • All nine blocks are present and ordered correctly (Problem, Customer Segments, UVP, Solution, Channels, Revenue Streams, Cost Structure, Key Metrics, Unfair Advantage)
  • Problem lists the top 3 problems AND includes an Existing Alternatives subsection
  • Customer Segments names Early Adopters as a distinct subset, not a restatement
  • UVP is one sentence AND includes a High-Level Concept analogy ("X for Y")
  • Solution maps 1:1 to the Problem block (top 3 features for top 3 problems)
  • Channels distinguishes compounding from traction-demonstrating paths
  • Revenue Streams shows the math (model, price, volume, LTV)
  • Cost Structure names CAC and identifies the cost driver
  • Key Metrics lists 3 to 5 leading indicators
  • Unfair Advantage is either specific OR explicitly flagged as an open question
  • Each block has a confidence tag (High|Medium|Low) with a one-line rationale
  • Evidence & Confidence section has all four subsections populated
  • Template guidance blockquotes are removed from the final artifact
  • Visual mode: the generated .html file opens successfully in a browser with no network access; every {{PLACEHOLDER}} has been replaced; confidence badges render with correct color; the canonical Maurya nine-block arrangement is visible; print-preview at A3 landscape fits the canvas on one page; role="img" and aria-label are present; the file path is reported back to the user
  • Content mode: no .html file is written and no visual section appears in the markdown

Examples

See references/EXAMPLE.md for a completed lean canvas in content mode with a realistic B2B SaaS scenario. The HTML template scaffold for visual mode lives at references/html-template.html.

平台分布

Codex

38.33%
按下载量换算31

Claude

29.63%
按下载量换算24

Cursor

19.25%
按下载量换算15

Gemini CLI

10.16%
按下载量换算8

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

操作浏览器

该 Skill 可能涉及浏览器控制能力,使用时可能读取或操作网页内容,需要在受控环境中确认权限边界。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills