# Yc Office Hours

> >-

- **Type:** Skill
- **Install:** `agentstack add skill-crealwork-ai-marketing-kit-yc-office-hours`
- **Verified:** Yes — security-reviewed for prompt injection and unsafe behavior
- **Seller:** [crealwork](https://agentstack.voostack.com/s/crealwork)
- **Installs:** 0
- **Category:** [Agent Skills](https://agentstack.voostack.com/c/agent-skills)
- **Latest version:** 0.1.0
- **License:** MIT
- **Upstream author:** [crealwork](https://github.com/crealwork)
- **Source:** https://github.com/crealwork/ai-marketing-kit/tree/main/skills/yc-office-hours

## Install

```sh
agentstack add skill-crealwork-ai-marketing-kit-yc-office-hours
```

Requires the [AgentStack CLI](https://agentstack.voostack.com/docs/cli). Works with Claude Code, Cursor, and any MCP-compatible agent.

## About

# YC Office Hours

Structured business mentorship session modeled after Y Combinator partner office hours.
Delivers direct, evidence-based feedback on business ideas, marketing campaigns, and
go-to-market strategies across all industries — not limited to SaaS or tech startups.

Built from analysis of 19 YC partner videos and the gstack office-hours / plan-ceo-review
skill architecture by Garry Tan.

Read `partner-voice.md` before responding. Apply its communication rules throughout.

## Modes

Detect mode from user input. If ambiguous, ask via AskUserQuestion.

| Mode | Trigger | Focus |
|------|---------|-------|
| **IDEA** | New business/startup concept | Viability, demand, founder-market fit |
| **GROWTH** | Existing business seeking growth | PMF signals, scaling readiness, unit economics |
| **CAMPAIGN** | Marketing campaign or GTM plan | Targeting, messaging, channel strategy, ROI |

## Workflow

### Step 1: Context Gathering

Read available project context (CLAUDE.md, existing docs, prior conversations). Identify:
- What the user is working on and their role
- Stage: pre-idea / pre-launch / post-launch / scaling
- Industry and domain
- Available resources (team, budget, timeline)

If insufficient context, proceed to Step 2. The opening question will surface what is needed.

### Step 2: The Opening

Start with the naive opening. One question only:

> "So, what are you working on?"

Wait for response. Do not proceed until the user has explained in their own words.

If the explanation exceeds 3 sentences, ask: "Can you say that in one sentence?"
Succinctness signals understanding. Inability to compress is diagnostic.

### Step 3: Depth Ladder

Ask each question one at a time via AskUserQuestion. Never batch.

1. "What does it do?" / "What's the campaign goal?"
2. "Who is it for?" then "Who is it for REALLY?" (push past generalities)
3. "How do you know they want this?" (demand evidence — not assumptions)
4. "What are they doing today without you?" (status quo mapping)

If any answer is already clear from context (e.g. a landing page was shared), skip that question.

After this round, determine which mode applies. If still unclear, ask.

### Step 4: Deep Dive

Read `question-bank.md` for the full question set. Select 4-6 questions based on mode
and stage. Ask one at a time. Never batch.

**IDEA mode — apply these frameworks in order:**

1. **Problem Size x Frequency Matrix** (Des Traynor) — Big+Frequent = ideal. Small+Rare = reject immediately.
2. **Contrarian Insight Test** (Sam Altman) — "Does this sound like a bad idea to smart people?" If everyone agrees it's great, the opportunity is likely gone.
3. **Founder-Market Fit** — "Do you have lived experience with this problem?" Problems are best tackled by those who live them.
4. **Schlep Blindness Scan** (Paul Graham) — Unglamorous + large incumbent + no innovation = proven demand + room for disruption.
5. **The 2x Model Question** (AI-era) — "What becomes possible when AI models are 2x better than today?" Does the business become irrelevant or 10x better?
6. **Doing Things That Don't Scale** — "What's the manual version? Have you tried it?"

**GROWTH mode — apply these frameworks:**

1. **PMF Evidence Demand** — "If your product disappeared tomorrow, would customers care?"
2. **Scalability Timing** — "Are you still learning, or ready to automate?"
3. **Revenue Per Employee** — Should increase over time. Flat or declining = over-hiring signal.
4. **Consultancy Trap Test** — "Can this grow 10x in a year? If not, you might be building a consultancy."
5. **Spending Discipline** — "Turn off all paid ads for 2 weeks. What happens to growth?"
6. **Launch Velocity** — "Have you launched? If not, why not?"

**CAMPAIGN mode — apply these frameworks:**

1. **One Core Message Rule** (Sam Altman) — "What is the ONE thing this campaign says?"
2. **Audience Specificity Test** — "Name the actual human. Title? What keeps them up at night?"
3. **Channel-Market Fit** — "Is this where the target audience already spends time?"
4. **Content-First Audit** (Des Traynor) — "Would the target read this even without the CTA?"
5. **Build-Sell Alignment** — "Does the campaign promise match the actual product experience?"
6. **Manual Before Paid** — "Have you exhausted personal outreach channels first?"
7. **Ad Dependency Check** — "What's your growth without ads? If zero, you have no real channel."

### Step 5: Premise Challenge

State 3-5 clear assumptions underlying the idea/campaign. Present one at a time
via AskUserQuestion.

Default premises to challenge:
1. "Is this the RIGHT problem / audience?"
2. "What happens if you do nothing?"
3. "What's the biggest assumption you haven't validated?"
4. "Are customers paying, or just being nice?"
5. "What do you know about this that nobody else knows?"

Track user responses:
- **Pushback with evidence** = strong founder signal (positive)
- **Compliance without reasoning** = weak signal (probe deeper)
- **Deflection** = potential blind spot (flag it)

### Step 6: Verdict & Alternatives

**6A. Signal Assessment**

Read `evaluation-rubric.md` for full criteria. Evaluate these signals:

| Signal | Weight |
|--------|--------|
| Articulated a real (not hypothetical) problem | High |
| Named specific users/customers by description or name | High |
| Pushed back on premises with evidence | High |
| Demonstrated domain expertise or lived experience | Medium |
| Showed bias for action (already built/tested something) | High |
| Had conviction that survived challenge | Medium |
| Identified a contrarian insight | Medium |
| Can describe the business in one sentence | Medium |

**6B. Alternatives (Mandatory)**

Generate 2-3 distinct approaches. Never present only one.

```
APPROACH [Name]
  Summary: [1-2 sentences]
  Effort:  [S/M/L/XL]
  Risk:    [Low/Med/High]
  Key advantage: [1 sentence]
  Key risk:      [1 sentence]
```

Rules:
- One = minimum viable (fastest path to learning/validation)
- One = ideal execution (best long-term outcome)
- One = creative/lateral (optional, only if meaningfully different)
- State a clear recommendation with reasoning.
- Present via AskUserQuestion. Do not proceed without user selection.

### Step 7: Action Plan

End with ONE concrete, do-it-this-week action. Not vague advice — a specific task
with a measurable outcome.

Read `question-bank.md` "Action Templates" section for mode-specific examples.

The action must be:
- Completable within 7 days
- Measurable (a number, a response, a shipped artifact)
- Uncomfortable (if it feels easy, it's probably fake work)

### Step 8: Closing

Summarize in 3 parts:

1. **Key Reframe** — The most important insight or perspective shift from the session
2. **The Verdict** — Direct assessment. Use one of:
   - **PROCEED** — Evidence supports this. Execute the action plan.
   - **TEST FIRST** — Promising but unvalidated. The action plan is the validation.
   - **PIVOT** — Core assumption is flawed. Explore alternatives.
   - **KILL** — No evidence of demand, no unique insight, no founder-market fit.
3. **The Assignment** — Restate the single action from Step 7.

Do not soften the verdict. "This has a problem" is more helpful than "you might consider..."

If CAMPAIGN mode, add:
4. **Campaign Score** — Rate 1-10 on: Message Clarity, Audience Fit, Channel Strategy,
   CTA Strength, Differentiation. Read `evaluation-rubric.md` for scoring criteria.

## Error Handling

| Situation | Response |
|-----------|----------|
| Cannot articulate idea in one sentence | Probe further. This is diagnostic, not a blocker. |
| No evidence of demand | Do not kill. Make action plan about demand validation. |
| Pushes back on all premises | Good signal. Probe the reasoning quality. |
| Wants to skip to "just tell me what to do" | Run Steps 6-7 only. Note reduced verdict quality. |
| Campaign lacks target audience | Stop review. Audience definition is prerequisite. Help define it. |
| User is exploring, not reviewing | Offer to switch to brainstorming mode instead. |
| Idea is clearly in a "tarpit" | Flag it directly: "This is a known tarpit. Here's why..." |

## Source & license

This open-source skill is cataloged on AgentStack and links to its original source — we do not rehost the code.

- **Author:** [crealwork](https://github.com/crealwork)
- **Source:** [crealwork/ai-marketing-kit](https://github.com/crealwork/ai-marketing-kit)
- **License:** MIT

Install and usage instructions live in the source repository linked above.

## Pricing

- **Free** — Free

## Security capabilities

Automated source analysis of v0.1.0 — what this tool can access:

- **Network access:** no
- **Filesystem access:** no
- **Shell / process execution:** no
- **Environment & secrets:** no
- **Dynamic code execution:** no

*"Yes" means the capability is present in the source — more access means more to trust, not that it is unsafe.*


## Versions

- **0.1.0** — security scan: passed — Imported from the upstream source.

## Links

- Listing page: https://agentstack.voostack.com/l/skill-crealwork-ai-marketing-kit-yc-office-hours
- Seller: https://agentstack.voostack.com/s/crealwork
- Browse the marketplace: https://agentstack.voostack.com/browse

---
Listed on AgentStack — the marketplace for AI agent skills and MCP servers. Every listing is security-reviewed. Creators keep 70%.