Token导航 LogoToken导航

Agent Skill · 研究检索

cs-onboard(船上 CS)

cs-onboard 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。

cs-onboard 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,权限为只读;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 41,189。

总安装 1,211周安装 49GitHub Stars 103下载 380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:cs-onboard(船上 CS)
来源仓库:https://github.com/borghei/claude-skills
仓库路径:skills/cs-onboard
安装命令:
npx skills add https://github.com/borghei/claude-skills --skill cs-onboard
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/borghei/claude-skills --skill cs-onboard

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。

cs-onboard 捕获 C 级高管所需的公司上下文,使后续建议更具针对性。适合在 Codex、Claude、Cursor、Gemini CLI 中进行创始人访谈与顾问接入前的背景整理。通过结构化对话生成持久化上下文文件,支撑定制化战略输出。

安装命令:npx skills add https://github.com/borghei/claude-skills --skill cs-onboard。使用前需确认权限范围、维护状态及是否会触发录音、存储或网络访问操作。

来源仓库内路径:skills/cs-onboard

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/borghei/claude-skills --skill cs-onboard;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

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

cs-onboard 怎么安装?

在终端执行 npx skills add https://github.com/borghei/claude-skills --skill cs-onboard。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

cs-onboard 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。

cs-onboard 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

cs-onboard 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。

cs-onboard 用的人多吗?

公开数据是累计安装 1,211 次,近一周安装 49 次,下载 380 次,GitHub 103 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。

cs-onboard 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。
  • 标注为只读权限,试用阶段不容易误改本地文件。

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。
  • 只想先试试效果、不希望它改动本地文件的人。

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
cs-onboard

船上 CS · 本页

1,211103只读未明确通过

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

C-Suite Onboarding

Tier: POWERFUL Category: C-Level Advisory Tags: onboarding, company context, founder interview, advisor setup, context capture, executive onboarding

Overview

C-Suite Onboarding captures the company context that powers every C-level advisory skill. One structured conversation produces a persistent context file that transforms generic advice into specific, situationally-aware guidance. Without context, advisory skills give textbook answers. With context, they give your-company answers.


Commands

CommandDurationPurpose
/cs:setup45 minutesFull onboarding interview across 7 dimensions
/cs:update15 minutesQuarterly refresh -- what changed since last capture
/cs:score5 minutesAssess context completeness and freshness

Interview Principles

This is a conversation, not a form. Follow these rules:

  1. One question at a time. Never list multiple questions.
  2. Follow threads. When something interesting surfaces, go deeper before moving on.
  3. Reflect back. "So the real issue sounds like X -- is that right?"
  4. Watch for what they skip. Avoidance signals the most important areas.
  5. Never read from a list. Use the framework as a map, not a script.
  6. Earn the hard questions. Start easy, build trust, then ask about weaknesses and fears.

Opening line:

"Tell me about the company in your own words -- what are you building and why does it matter?"

Then let the conversation flow naturally across the 7 dimensions.


The 7 Interview Dimensions

Dimension 1: Company Identity

What to capture: What they do, who it is for, the real founding "why," one-sentence pitch, non-negotiable values.

Key probes:

  • "If you had to explain this to your grandmother in one sentence, what would you say?"
  • "What is a value you would fire someone over violating?"
  • "What do you refuse to compromise on, even if it costs you?"

Red flags:

  • Values that sound like marketing copy ("we empower synergies")
  • Cannot articulate the founding "why" beyond "I saw a market opportunity"
  • Mission statement that could apply to any company in the industry

Dimension 2: Stage & Scale

What to capture: Headcount (FT vs contractors), revenue range, runway, stage label, what broke in the last 90 days.

Key probes:

  • "If you had to label your stage -- still finding PMF, scaling what works, or optimizing -- which is it?"
  • "What broke in the last 90 days that you did not expect?"
  • "What is the one metric you check every morning?"

Stage Definitions:

StageSignalTypical Challenges
Pre-PMF<$500K ARR, pivoting, searchingFinding the right customer/problem/solution fit
Early PMF$500K-$2M ARR, repeatable salesHiring, process, not breaking what works
Scaling$2M-$10M ARR, growth machine buildingMiddle management, culture preservation, unit economics
Optimizing$10M+ ARR, efficiency focusInnovation stagnation, org complexity, market defense

Dimension 3: Founder Profile

What to capture: Self-identified superpower, acknowledged blind spots, archetype, what actually keeps them up at night.

Key probes:

  • "What would your co-founder (or closest advisor) say you should stop doing?"
  • "When things go wrong, what is your instinctive first reaction?"
  • "What part of running this company do you secretly dislike?"

Founder Archetypes:

ArchetypeStrengthBlind SpotAdvisory Focus
ProductDeep user empathy, product visionSales, go-to-market, delegationRevenue strategy, hiring
TechnicalEngineering excellence, technical moatsBusiness model, communicationGTM, storytelling, team
SalesRevenue generation, relationshipsProduct depth, technical debtProduct strategy, engineering
OperatorExecution, processes, efficiencyVision, innovation, risk-takingStrategy, long-term planning

Red flags:

  • No acknowledged blind spots
  • Weakness framed as strength ("I'm too much of a perfectionist")
  • Co-founder dynamics described as "fine" with no specifics

Dimension 4: Team & Culture

What to capture: Team described in 3 words, last real conflict and how it was resolved, which values are real vs aspirational, strongest and weakest leader.

Key probes:

  • "Which of your stated values is most real? Which is a poster on the wall?"
  • "Tell me about the last real disagreement in the leadership team."
  • "Who is the one person you cannot afford to lose, and why?"

Red flags:

  • "We have no conflict" -- every healthy team has conflict
  • Cannot name a weak leader -- either lying or not paying attention
  • Culture described only in positive terms with no self-awareness

Dimension 5: Market & Competition

What to capture: Who is winning and why (honest version), real unfair advantage, the competitive move that could hurt them.

Key probes:

  • "What is your real unfair advantage -- not the investor pitch version?"
  • "If your best competitor had unlimited funding, what would they do that scares you?"
  • "What do your competitors do better than you? Be honest."

Red flags:

  • "We have no real competition" -- you always have competition, even if it is the status quo
  • Unfair advantage is a feature that can be copied in 6 months
  • No awareness of competitor strategy

Dimension 6: Current Challenges

What to capture: Priority stack-rank across product/growth/people/money/operations, the decision they have been avoiding, the "one extra day" answer.

Key probes:

  • "If you had one extra day per week, what would you spend it on?" (Reveals true priority)
  • "What is the decision you have been putting off for weeks?"
  • "Rank these: product, growth, people, money, operations. What is number 1 right now?"

The "avoided decision" is often the most valuable insight. Common avoided decisions:

  • Firing an underperformer who is well-liked
  • Pivoting away from a product that is not working
  • Having an honest conversation with a co-founder
  • Raising prices
  • Cutting a feature or initiative

Dimension 7: Goals & Ambition

What to capture: 12-month target (specific and measurable), 36-month target (directional), exit vs build-forever orientation, personal success definition.

Key probes:

  • "What does 12 months from now look like if everything goes right?"
  • "What does success look like for you personally -- separate from the company?"
  • "Are you building to sell, building forever, or haven't decided?"

Red flags:

  • 12-month target that is vague ("grow a lot")
  • Personal and company goals completely disconnected
  • Exit orientation that conflicts with stated mission

Context Output Structure

After the interview, generate the context file at ~/.claude/company-context.md:

# Company Context
**Last updated:** [Date]
**Freshness:** Fresh (< 90 days)
**Completeness:** [X/7 dimensions captured]
**Interviewed:** [Founder name and role]

## 1. Company Identity
**What we do:** [One sentence]
**Who it's for:** [Target customer]
**Why it matters:** [Founding motivation -- the real version]
**One-line pitch:** [Elevator pitch]
**Non-negotiable values:** [Values they would fire over]

## 2. Stage & Scale
**Stage:** [Pre-PMF / Early PMF / Scaling / Optimizing]
**Headcount:** [X FT + Y contractors]
**Revenue:** [$X ARR / MRR]
**Runway:** [X months at current burn]
**Morning metric:** [What they check first]
**Recent break:** [What broke in last 90 days]

## 3. Founder Profile
**Archetype:** [Product / Technical / Sales / Operator]
**Superpower:** [Self-identified strength]
**Blind spot:** [Acknowledged weakness]
**Up-at-night:** [Current anxiety]
**Co-founder dynamic:** [Healthy / Strained / Solo]

## 4. Team & Culture
**Team in 3 words:** [Their words]
**Real values:** [Values that are actually lived]
**Aspirational values:** [Values that are work-in-progress]
**Key person risk:** [Who they cannot lose]
**Weakest link:** [Leadership gap]
**Last conflict:** [What happened and how resolved]

## 5. Market & Competition
**Winning competitor:** [Who and why]
**Real unfair advantage:** [Not the investor version]
**Kill-shot risk:** [Competitive move that could hurt them]
**Market position:** [Leader / Challenger / Niche / Emerging]

## 6. Current Challenges
**Priority #1:** [Top challenge area]
**Priority stack:** [Rank of product/growth/people/money/ops]
**Avoided decision:** [What they have been putting off]
**One-extra-day:** [What they would spend time on]

## 7. Goals & Ambition
**12-month target:** [Specific, measurable]
**36-month target:** [Directional]
**Orientation:** [Build to sell / Build forever / Undecided]
**Personal success:** [What success means for the founder personally]

## Notes
[Observations, inferred patterns, things to watch]
[CONTEXT UPDATE entries from subsequent sessions]

Rules:

  • Write [not captured] for unknowns -- never leave blank
  • Use their actual words when possible, not corporate paraphrasing
  • Note confidence level for inferred information
  • Never silently modify -- always confirm before updating

Context Quality Scoring

Completeness Score

Dimensions CapturedScoreLabel
7/7100%Complete
5-6/770-85%Good (identify gaps)
3-4/740-55%Partial (schedule follow-up)
1-2/715-25%Minimal (re-interview needed)

Freshness Score

AgeScoreLabelAction
< 30 daysFreshHigh confidenceUse directly
30-90 daysAgingMedium confidenceUse, flag what may have changed
90-180 daysStaleLow confidencePrompt for /cs:update
> 180 daysExpiredVery low confidenceRe-interview recommended

Quality Signals

SignalConfidence Impact
Full interview completed+High
Update done within 90 days+Medium
Key fields populated with specifics+High
Fields contain vague/generic answers-Medium
Financial fields missing-High (for CFO/CEO skills)
"Not captured" in 3+ fields-High

Quarterly Refresh Protocol (/cs:update)

Trigger: Every 90 days or after a major event (fundraise, reorg, pivot, key hire/departure).

Opening: "It has been [X time] since we captured your company context. Let's do a quick refresh. What has changed?"

Walk each dimension with a single "what changed?" question:

DimensionRefresh Question
Identity"Still the same mission, or has it shifted?"
Stage & Scale"Team size, revenue, and runway now?"
Founder"Has your role changed? What is stretching you?"
Team"Any leadership changes? New key players?"
Market"Any competitive surprises? Market shifts?"
Challenges"What is the #1 problem now vs 90 days ago?"
Goals"Still on track for the 12-month target?"

After refresh:

  • Update relevant sections in context file
  • Update Last updated timestamp
  • Reset freshness to Fresh
  • Note what changed in the Notes section

Context Enrichment During Sessions

During advisory conversations, new information surfaces. Capture it without disrupting flow.

Triggers for enrichment:

  • New metric or number shared
  • Key person mentioned for the first time
  • Priority shift expressed
  • New constraint or risk surfaces
  • Timeline or deadline revealed

Protocol:

  1. Note internally during conversation
  2. At session end: "I picked up a few things that would be useful to add to your context. Want me to update the file?"
  3. If yes: append to relevant section, update timestamp
  4. If no: respect the decision

Never silently modify the context file. Always confirm before changes.


Privacy Rules

Never Send Externally

  • Specific revenue or burn rate figures
  • Customer names (unless publicly known)
  • Employee names (unless publicly known)
  • Investor names (unless publicly announced)
  • Specific runway months
  • Watch list contents
  • Avoided decisions

Safe to Use Externally (Anonymized)

  • Stage label (seed, Series A, etc.)
  • Team size ranges (1-10, 10-50, 50-200+)
  • Industry vertical
  • Challenge category (not specifics)
  • Market position descriptor

Before Any External API Call

  • Numbers become ranges or stage-relative descriptors
  • Names become roles ("the CTO" not "Sarah")
  • Revenue becomes stage labels ("early revenue" not "$800K ARR")
  • Customers become "Customer A, B, C"

Missing Context Handling

Handle gracefully -- never block the conversation.

MissingApproach
Stage"Just to calibrate -- are you still finding PMF or scaling what works?"
FinancialsUse stage + team size to infer. Note the inference.
Founder profileInfer from conversation style. Mark as inferred.
Multiple foundersContext reflects interviewee. Note co-founder perspective may differ.
Recent context"It's been a while since your last update. Should I assume things are roughly the same, or has something big changed?"

Integration with C-Suite Skills

The context file is loaded by every C-level advisory skill:

SkillUses Context For
ceo-advisorStrategic recommendations calibrated to stage and challenges
cfo-advisorFinancial guidance calibrated to runway and revenue
cto-advisorTechnical strategy calibrated to team and architecture
coo-advisorOperations advice calibrated to scale and processes
cmo-advisorMarketing strategy calibrated to stage and market position
internal-narrativeNarrative construction based on company truth
scenario-war-roomRisk variables calibrated to actual threats

Related Skills

SkillUse When
ceo-advisorUsing context for strategic decisions
internal-narrativeBuilding narratives from the captured context
scenario-war-roomModeling risks based on company context
context-engineTechnical implementation of context management for AI agents

Troubleshooting

ProblemLikely CauseResolution
Founder gives vague, marketing-speak answersInterview not building trust; questions too formalReset with a personal question; follow a thread they care about; use reflection ("So what you're really saying is...")
Key dimensions left as "[not captured]" after interviewFounder avoided topic or interview ran out of timeNote which dimensions were skipped (avoidance signals importance); schedule targeted follow-up for missing areas
Context file feels generic, could apply to any companyInterviewer used the framework as a script, not a mapRe-interview with focus on specifics: names, numbers, stories; capture their actual words, not paraphrases
Advisory skills still giving textbook answers despite contextContext file too sparse or stale (>90 days)Run /cs:score to assess completeness and freshness; schedule /cs:update if freshness is Stale or Expired
Quarterly refresh takes too long or feels redundantRefresh covering all 7 dimensions instead of focusing on changesLead with "What changed?" per dimension; skip unchanged areas quickly; target 15 minutes total
Multiple founders give conflicting contextDifferent perspectives on stage, challenges, or prioritiesNote the conflict explicitly in context file; mark which founder provided which data; flag for resolution
Context file modified without founder approvalAI agent or team member updated context silentlyEnforce "never silently modify" rule; require explicit founder confirmation before any context update

Success Criteria

  • Context completeness score reaches 7/7 dimensions within first session or first follow-up
  • Context freshness maintained at "Fresh" (< 90 days old) through quarterly refresh protocol
  • Advisory skills produce company-specific (not generic) recommendations when context is loaded
  • Founder reports that AI advisory conversations feel "like talking to someone who knows my business"
  • Time to complete full onboarding interview under 45 minutes
  • Time to complete quarterly refresh under 15 minutes
  • Zero instances of context file modified without founder approval

Scope & Limitations

In scope: Structured founder interview across 7 dimensions (identity, stage, founder profile, team/culture, market/competition, challenges, goals), context file generation and maintenance, quarterly refresh protocol, context quality scoring (completeness and freshness), context enrichment during advisory sessions, privacy rules for external data handling, and missing context graceful handling.

Out of scope: Company financial modeling (use cfo-advisor), competitive intelligence gathering (use competitive-intel), team assessment or 360 reviews (use hr-operations/), product analytics (use cpo-advisor), and automated context capture from external data sources. This skill captures context through conversation, not data integration.

Limitations: Context quality depends entirely on founder candor; evasive or aspirational answers reduce advisory effectiveness. Single-founder interviews may miss co-founder perspectives. Context is a point-in-time snapshot; rapid company changes (pivot, reorg, fundraise) can make context stale before the 90-day refresh. Privacy rules prevent sharing specific context externally, which limits integration with external tools.


Integration Points

  • ceo-advisor -- Strategic recommendations calibrated to company stage, challenges, and founder archetype
  • cfo-advisor -- Financial guidance calibrated to runway, revenue range, and growth stage
  • cto-advisor -- Technical strategy calibrated to team size, architecture maturity, and founder technical depth
  • coo-advisor -- Operations advice calibrated to scale, process maturity, and headcount
  • cmo-advisor -- Marketing strategy calibrated to stage, market position, and competitive landscape
  • internal-narrative -- Narrative construction uses company identity, values, and founder voice from context
  • scenario-war-room -- Risk variables calibrated to actual competitive threats and company vulnerabilities

平台分布

Codex

35.01%
按下载量换算133

Claude

29.09%
按下载量换算111

Cursor

18.82%
按下载量换算72

Gemini CLI

8.71%
按下载量换算33

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

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