Product Market Fit: AI Agent Templates

The Marc Andreessen Agent with Claude

Guillermo Flor — Sat, 23 May 2026 12:34:24 GMT

Imagine having Marc Andreessen on speed dial.

Ask him whether to raise now or wait. Whether to rebuild for AI or add features. Whether the big tech competitor that just announced your product is a threat or a validation. Most founders will never get that access. But with Claude, you can get surprisingly close.

Marc has published more than almost any investor alive. 20 years of thinking on markets, startups and technology, all sitting in public, mostly unread.

This guide teaches you how to scrape it all into one folder, load it into Claude Code with a system prompt that grounds every response in his actual writing and interrogate it like an advisor. Not a chatbot guessing what Marc sounds like. An agent that cites the source file before it answers.

📚 By the end you’ll have the scraper, the system prompt, and real examples of founder questions with Marc-style outputs grounded in his actual catalog.

Inside you will find:

Why Marc’s Public Catalog Is More Useful Than a VC Call
How to Scrape Five Sources Into One Folder With Claude Code
The System Prompt That Makes Claude Respond Like Marc
Which Questions the Agent Answers Best
Five Real Founder Questions

👋 Just a quick note before you continue!

Building great AI systems and agents is quickly becoming one of the most important advantages a startup can have. So we’ve been building non-stop and this will become a much bigger part of PMF going forward.

From now on, as a premium subscriber, you’ll get access not just to the best resources on fundraising and growth (160+ Playbooks, 65+ Success Stories, 30+ Pitch Deck Collections, 25+ Investor Resources), but also:

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

Let’s begin!

How the Fastest Teams Actually Ship Code with AI

Guillermo Flor — Fri, 22 May 2026 15:15:06 GMT

Welcome to Product Market Fit! Today’s issue is free and open to everyone.

This week we host Paul Sanglé-Ferrière, CEO of cubic (YC). His team has built the leading AI code review tool, ranked #1 on every independent benchmark and beating tools from Anthropic and OpenAI by a large margin. They build for companies like Cloudflare, n8n, and Legora and outship teams that have raised 100x more. Paul is one of the world’s leading experts on harness engineering and AI coding workflows.

P.S. At the end, premium subscribers will find an exclusive discount to try cubic!

Today, I want to show you how the fastest engineering teams I know actually ship code today.

Writing code used to be the bottleneck. With AI, that’s stopped being true. Teams are shipping ten or twenty times more code than they were a year ago, and review is what can’t keep up. Senior engineers don’t have time to read every PR properly, so bugs slip through and team standards quietly erode. The faster you ship, the harder it gets to know what’s actually in your own codebase.

cubic is the AI code reviewer the most advanced engineering teams in the world rely on.

To be concrete: cubic doesn’t write code. It’s a specialised layer that reviews it.

If you’re writing code for production at scale, you need a dedicated reviewer. Code-writing agents have their own blind spots, and they’re not built to review their own work. You need something that understands your team, runs different models from your coding agent, and is purpose-built for review.

You still use your coding agent of choice (Claude, Codex, Cursor) for the writing. cubic is what runs after, replacing whatever you were using to review AI-written code before, whether that was you eyeballing the diff or another tool.

Internally, we’re a team of five. We outperform teams that have raised 100x more than we have. Our reviewer ranks #1 on every independent code review benchmark, beating every other competitor by a wide margin.

Between the two, the same workflow keeps showing up. We see how the best engineering teams in the world ship, and we ship the same way ourselves.

📚 That's what I want to walk through here. The loop, roughly: set up your repo so your agent works the way you do, plan before you prompt, let the agent write, have a different agent review, ship and let a deeper agent review run overnight.

Inside you will find:

Before You Prompt Anything
Planning Before Coding
Letting the Agent Build
The Review Loop
In the PR, Go Deeper
Overnight Review
The Full Loop
BONUS: exclusive discount to try cubic!

Let’s start at the beginning.

1. Before You Prompt Anything

The goal at this stage is simple: give your agent enough context that it doesn’t have to guess. Your repo layout, team conventions, standards, all in a place it can read once and refer back to.

Get this right and every prompt downstream gets shorter and better. Skip it and you’ll be re-explaining yourself in every conversation.

This is the boring part, but it matters.

Put an AGENTS.md (or CLAUDE.md) at the root of your repo. Mine has:

Repo layout, the directories that actually matter
Build, test, and lint commands
Engineering conventions we follow
A short list of “do not do this” rules
What “done” actually looks like for our codebase

Your agent reads it once and stops asking you the same questions in every prompt.

One thing worth doing on top of that: hook up the cubic MCP to your coding agent.

cubic learns from how your senior engineers review code. With the MCP connected, your agent already knows what your team cares about before it writes a line. “What would the senior reviewer flag here?” becomes something the agent can answer, not guess at.

2. Planning Before Coding

Use Plan mode. In Cursor that’s Shift+Tab, or /plan.

Let the agent interview you first. The temptation is to drop one giant prompt in and hope. Resist it. A five-minute back-and-forth gets you better output than a perfectly worded one-shot.

One thing that’s underrated: dictation. There are a bunch of tools for it, I use Willow Voice. You talk a lot faster than you type, so you can dump way more context into the plan and actually have a high-speed back-and-forth with the agent instead of laboriously typing each turn.

Break work into thin slices: plan, approve, implement, test, cleanup. Each one’s a small conversation.

A few less obvious things I do here.

One of the biggest weaknesses of today’s AI agents, even the best ones, is design. They’re bad at it. Drop a design problem on them cold and the output usually looks like every other AI-generated UI on the internet right now.

The thing they are good at is copying. Give them strong references and they perform much better.

So for any UI work, I solve the design problem before I solve the code problem. I point an MCP like Mobbin at the plan so the agent has real reference designs to pull from instead of guessing. Then I have it sketch low-fidelity ASCII diagrams of the layout before any real code gets written. Way easier to argue about structure on a wireframe than in a 600-line diff.

The last one isn’t design-specific: match the model to the task. Big model for architecture, cheaper model for the grind.

3. Letting the Agent Build

Once the plan is in place, you can let the agent run most of the loop itself. Models are good enough now to self-debug.

Have it run your app locally, see what breaks, and iterate until the build is green and the tests pass. You don’t need to sit there feeding it each compile error by hand.

If there’s a UI involved, ask the agent to take screenshots of what it’s built as it goes. It’s surprisingly good at catching its own mistakes when it can actually see the output instead of just diffing text.

When it says it’s done, do one last check: does the output actually respect the ASCII diagrams and references from the plan? Drift happens, especially in longer sessions. If it drifted, send it back and have it fix the gap.

4. The Review Loop

There are two real problems with shipping AI-written code.

The first is the obvious one: agents make mistakes. They hallucinate APIs. They write insecure patterns. They write queries that look fine and run for ten seconds in production. You can’t catch all of that in a 30-second diff skim, especially when you’re shipping ten of these a day.
The second is more subtle: agents don’t know your context. They don’t know you’re on an MVP where speed matters more than perfection, or that this service is a critical path that can’t go down. They don’t know the standards your team has spent years settling on.

This is where cubic comes in. You don’t run the review yourself. You tell your coding agent to call the cubic CLI when it’s done writing, and the two agents iterate together until there’s nothing left to flag. A different agent (different models, different blind spots) reviews the diff, hands the issues back to your coding agent, and your coding agent fixes them. Fully agentic. You don’t sit there driving it.

I think the different-model bit is the most underrated part of all this. If the same model that wrote the code also reviews it, you’ll miss the same bugs twice. The blind spots are identical. You want a different reviewer on the other side.

cubic actually does this for you automatically. We detect which model wrote the code and route the review to a different one. If we see Opus generated the diff, we’ll review it with Codex. If it was Codex, we’ll review with Opus. You don’t have to think about it.

But the part that really moves the needle is that cubic knows your team. It syncs your standards, the patterns your senior engineers actually flag in review, the bits specific to your repo. You don’t get a generic review, you get one that reflects what your team actually cares about.

By the time the loop finishes, the diff is already cleaned up. You step in to approve or push back, not to do the review yourself.

5. In the PR, Go Deeper

Push to GitHub.

cubic’s GitHub agents do a heavier pass than the CLI. More context, more thinking time.

Use the cubic skill in your IDE to talk back to the comments. Same loop as before, just with a more thorough reviewer on the other end.

6. Overnight Review

Merge your PR. Ship it. Go home.

Once you merge, cubic kicks off a deeper review on whatever just landed in main. It runs for hours, up to twelve of them, doing codebase-level analysis. Cross-file logic, security holes, the kind of bug you only catch with full-repo context. Things a 90-second PR review just can’t surface.

If it finds something, there’s a fix PR waiting for you in the morning.

It’s the kind of compute budget you only get when nobody’s waiting on the answer.

7. The Full Loop

End to end, this is how we ship:

Setup: AGENTS.md, environment, cubic MCP connected
Plan: plan mode, let the agent interview you, thin slices
Write: coding agent does it
Local review: cubic CLI before you push
PR review: cubic GitHub agents go deeper
Merge
Overnight: cubic deep review runs while you sleep
Morning: merge the fix PR if there is one

Writing and reviewing are both jobs for agents now. What’s left for you is judgment, and the call on whether to actually ship.

If you take one thing from this, take this: don’t let the same model that wrote your code be the one signing off on it.

8. BONUS: exclusive discount to try cubic! (for Paid subscribers)

How to Use Gamma with Claude: The Playbook

Guillermo Flor — Thu, 21 May 2026 16:40:42 GMT

A pitch deck used to take a weekend in Keynote. A board memo, an afternoon in Notion. A launch page, days in Webflow or thousands at an agency. A LinkedIn carousel, hours in Figma.

Every founder communication artifact had its own tool, its own time cost and zero brand consistency between them.

Earlier this year Gamma shipped a native MCP connector inside Claude. The same four artifacts now get generated in seconds, in the same chat where you’re already thinking through the argument: themed, exported and ready to share.

📚 This playbook is the operational guide to using Gamma + Claude as your founder communication machine, from investor pitch to social carousels. By the end you’ll have the connector set up, six copy-paste prompts that cover 95% of what a founder produces, and a Claude Project that keeps every future output looking like the same company built it.

Inside you will find:

Setup: Connect Gamma to Claude and What You’ll Be Building
Pitch Deck: The Investor Prompt
Sales Deck: The Customer-Facing Prompt
Board Update: The Document Prompt
Launch Page: The Webpage Prompt
LinkedIn Carousel: The Social Prompt
How to Brand It All: Themes, Logos, Style Presets
The Claude Project That Ties It Together

👋 Just a quick note before you continue!

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

Here, few examples:

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

Let’s begin!

The Agent Builder’s Playbook: 8 Lessons from Anthropic

Guillermo Flor — Wed, 20 May 2026 16:10:46 GMT

Tariq, an engineer at Anthropic on the team that builds Claude Code, spent two hours at the AI Engineer Summit teaching how the company actually builds agents internally.

The session was dense and full of code, but underneath the syntax there were nine ideas you can apply today even if you’ve never written a tool call in your life.

📚 I analyzed the full session and pulled out the nine moments where Tariq said something that would change how a founder makes a decision today. Each one comes with a real example and a “do this” step.

At the bottom there’s a bonus: a starter system prompt that bakes the framework together, ready to adapt for your own agent.

Inside you will find:

Lesson 1: Build every agent around the Gather-Act-Verify loop
Lesson 2: Run every idea through the Verification Filter
Lesson 3: Apply the Reversibility Filter to know what to trust
Lesson 4: Pass the Locked-in-a-Room Test when designing tools
Lesson 5: Use the In-Distribution Trick to save hours
Lesson 6: Read the transcripts Every time
Lesson 7: Embrace the Throwaway Code advantage
Lesson 8: Price for the hard problem
Bonus: The Starter System Prompt

Here’s what I’d give to anyone building out their GTM motion for the first time: GTM Atlas.

Curated by Attio, the AI CRM, it’s the ultimate resource for early-stage teams navigating growth from scratch. You’ll get:

Systems thinking that scales with you
A point of view on your GTM, not borrowed from a playbook
Exclusive stories from industry leaders setting the standard

The first installation features entries from Elena Verna (Lovable), Kyle Norton (Owner.com), Roniesha Copeland (Vercel), James Pastan (Framer), and more.

Mapped by operators. Curated by Attio.

Explore Now!

👋 Just a quick note before you continue!

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

Here, few examples:

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

Let’s begin!

Kimi K2.6: The Ultimate Claude Alternative

Guillermo Flor — Tue, 19 May 2026 15:46:48 GMT

Moonshot, a Beijing-based AI lab backed by Alibaba, just released Kimi K2.6. It is the first open-source model that:

beats Claude Opus on the coding tasks closest to real engineering work
costs roughly one tenth of what Anthropic charges per token
it can run a single automated task for thirteen hours straight without falling apart

This is not another “cheap Claude alternative” that breaks the moment the task gets real. K2.6 ships with downloadable weights on Hugging Face, an official command-line tool that mirrors Claude Code and integration paths through every major inference provider. You can run it from your browser, from your terminal, or inside your own infrastructure if you have the GPUs.

📚 By the end of this guide you will know exactly what ships in Kimi K2.6, where it beats Claude and where it doesn’t, the real cost math for a founder, the four ways to run it, the 7 things Kimi can do for you on day one without writing a line of code and the routing rules that keep the rest of your stack honest.

Inside you will find:

What Actually Ships in Kimi K2.6
Where K2.6 Beats Claude
The Cost Math That Changes a Founder’s Monthly AI Bill
The 3 Ways to Run K2.6
The 7 Things Kimi Can Do For You Right Now (including OpenClaw)
The Routing Rule: When to Keep Claude in Your Stack

Launch fast. Design beautifully. Build your company’s website on Framer.

Framer helps teams design, build, and launch their marketing sites lightning fast.

With the ability to publish hundreds of CMS pages in a single click, operate at a global scale with seamless localization, and even host unified content across multiple domains, teams have never been able to ship faster.

Trusted by companies like Miro, Bilt, and Perplexity.

Speed without chaos: ship pages and updates faster without turning the site into a fragile set of one-off hacks
Reduce dependency: shift routine brand and marketing work out of product engineering queues.
Production-grade foundation: Run real marketing systems (CMS, SEO, performance optimization) with governance and collaboration

👉 Build your company’s site on Framer today

Apply to Claim Your Free Year

👋 Just a quick note before you continue!

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

Here, few examples:

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

Let’s begin!

Claude Agent for CRM Maintenance

Guillermo Flor — Mon, 18 May 2026 18:12:21 GMT

A B2B CRM loses between 22% and 70% of its accuracy every year, depending on the industry.

And the cost lands where it hurts: sales reps waste 27% of their time pursuing contacts that don’t exist anymore, somewhere north of five hundred hours per person per year, on calls that were never going to connect.

The manual fix doesn’t match the shape of the problem. Cleanup is bursty while decay is continuous.

That’s why we are going to build a Claude agent that scans every account every day.

📚 By the end of this guide you’ll have a working CRM maintenance agent running inside Claude Code: a project that scans every contact daily, detects job changes within twenty-four hours, validates each detection against evidence before touching a record and writes corrections back to your CRM through MCP.

Inside you will find:

How to Set Up the CRM Maintenance Agent
How to Detect Job Changes Before Your Competitors Do
How to Validate Every Detection Before It Hits the CRM
How to Refresh Contact Data Without Breaking the CRM
How to Turn Champion Moves Into Warm Pipeline
How to Keep Bad Records Out of Your Sequences
How to Run the Agent Every Day Across Your Full Database

PS: I’m partnering with Deel to fund up to 10 founders with $1,000,000 🤑

Apply now

👋 Just a quick note before you continue!

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

Here, few examples:

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

Let’s begin!

Claude for Small Business: The Complete Guide

Guillermo Flor — Sun, 17 May 2026 17:55:09 GMT

Small businesses are 44% of US GDP. They employ nearly half the private-sector workforce. And still they are the segment of the economy where AI adoption had visibly stalled.

The AI wave of the last two years was built for engineers and enterprise teams with a real stack and internal IT.

For the founder running a small business, AI stopped at the chat window: type a question, paste an answer. Useful but bounded. Meanwhile she still ran the books, the payroll, the campaigns, the customer emails, the contracts, and the hiring on her own.

Few days ago Anthropic shipped the thing that closes that gap. Claude for Small Business is a one-click plugin that ships with 15 ready-to-run agentic workflows and 15 underlying skills built on the repeatable tasks small business owners said slow them down most.

📚 This guide is the complete walkthrough. How the plugin is structured, how to install it, every skill explained with the exact prompt you can paste into Cowork tomorrow, three combined workflows that change your week, and how to customize the whole thing so it knows your business.

Inside you will find:

How to Install Claude for Small Business in 10 Minutes
Money and Finance: Five Workflows That Run Your Back Office
Sales, Marketing, and Customer Ops: Seven Workflows for the Front Line
Business Intelligence and Hiring: Six Skills That Sit Above the Day-to-Day
Three Combined Workflows That Change Your Week
How to Customize the Plugin So It Knows Your Business

👋 Just a quick note before you continue!

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

Here, few examples:

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

Let’s begin!

How to Build a Second Brain with Claude

Guillermo Flor — Thu, 14 May 2026 19:34:26 GMT

Scattered. Forgotten. Never connected.

That is what happens to every meaningful document a startup produces.

The customer call that explains the churn. The line in the financial model that contradicts the pitch. The pattern in the last six one-on-ones with the CTO. All stored, none ever opened again, none ever placed next to each other.

You already have everything you need. The problem is that nothing speaks to anything else.

Here is what we are building
One folder on your Desktop. Every important document routed into it. Claude reading across the entire thing on demand. You stop searching, you start asking. “What objection is killing deals this quarter?” “Where does my model contradict reality?” Twelve months from now, your second brain will have read more of your own work than you ever could.

Inside you will find:

How Your Best Thinking Already Slips Through Your Hands
How to Structure a Folder Claude Can Actually Reason Across
The 10-Second Capture Habit
How to Connect Claude to Your Second Brain
How to Get Claude to Find the Patterns Hiding in Your Own Work (15 Prompts)
How to Run the Weekly Review That Makes the System Compound

PS: I’m partnering with Deel to fund up to 10 founders with $1,000,000 🤑

Apply now

👋 Just a quick note before you continue!

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

Here, few examples:

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

Let’s begin!

The Ultimate Claude Code Repo (181k-Stars)

Guillermo Flor — Wed, 13 May 2026 18:43:20 GMT

A Claude Code repo just crossed 181,000 stars. Almost 28,000 forks. All of this in less than ten months.

Inside: 28 agents, 119 skills, and a 354-line guide on using Claude Code like a pro.

Most people who star it never go back. The README is a wall.

I read all of it for you. Inside the mass I picked the agents, skills, and tips most useful to a startup founder. Below a playbook on how to use each one, when it pays off, and the copy-paste prompt to actually run it.

Here’s the cut.

The 5 skills:

market-research: Source-cited research on competitors, markets, and investors.
investor-materials: Pitch decks, one-pagers, memos, and financial models in VC-ready formats.
investor-outreach: Personalized cold emails to investors plus the follow-up cadence.
article-writing: Long-form writing in your voice, without the typical AI tone.
content-engine: One asset into native posts across LinkedIn, X, and beyond.

The 5 agents:

chief-of-staff: Communication triage and ready-to-send drafts.
planner: Turns a vague idea into operative steps with dependencies.
docs-lookup: Answers questions buried inside external documentation.
code-reviewer: Reviews what AI wrote before you ship it.
security-reviewer: Catches vulnerabilities before they hit production.

The 5 tips:

Memory persistence → Three hooks that make Claude remember between sessions.
Continuous learning → Claude writes skills from its own mistakes.
Model routing → Haiku for cheap work, Sonnet by default, Opus only when needed.
Turn off the MCPs you don’t really need →They eat tokens for nothing.
The /llms.txt trick → LLM-ready versions of any documentation in one click.

Inside you will find:

The 5 business skills built explicitly for founders
The 5 pre-built Claude agents that replace categories of busy-work
The 5 tips to use Claude like a pro

Let’s begin!

👋 Just a quick note before you continue!

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

Here, few examples:

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

How to Build a Customer Support Startup on ElevenLabs

Guillermo Flor — Tue, 12 May 2026 18:35:33 GMT

The voice customer support category has been a stagnant pond for ten years.

That changed in 2025. ElevenLabs, Vapi, Retell, Bland shipped voice agents with sub-second turn-taking, expressive speech that crossed the perceptual threshold.

The opportunity that opens for founders it’s huge. And vertical.

The CCaaS (Customer Care as a service) giants serve enterprise. Mid-market companies and SMBs in specific verticals (dental clinics, law firms, e-commerce returns, real estate, plumbing, home services, B2B SaaS support) are entirely underserved.

📚 By the end of this guide you’ll have an operational playbook: a framework for picking the vertical you can actually win, the full build of the agent template that runs every customer’s traffic on ElevenLabs and the operating loop with Claude that lets one founder run 50 tenants without hiring.

Inside you will find:

Why Voice Customer Support Just Became a Vertical SaaS Opportunity
How to Pick the Vertical That Will Carry Your Startup
How to Design the Wedge That Makes You Hard to Copy
How to Build the Agent Template That Powers Every Customer
How to Onboard a New Customer and Deploy on Their Site, Phone, and WhatsApp
How to Operate Every Customer’s Agent With Claude
How to Price the Agent

Launch fast. Design beautifully. Build your company’s website on Framer.

Framer helps teams design, build, and launch their marketing sites lightning fast.

Trusted by companies like Miro, Bilt, and Perplexity.

Speed without chaos: ship pages and updates faster without turning the site into a fragile set of one-off hacks
Reduce dependency: shift routine brand and marketing work out of product engineering queues.
Production-grade foundation: Run real marketing systems (CMS, SEO, performance optimization) with governance and collaboration

👉 Build your company’s site on Framer today

Apply to Claim Your Free Year

👋 Just a quick note before you continue!

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

Here, few examples:

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

Let’s begin!

The 10 Prompting Hacks by Anthropic

Guillermo Flor — Mon, 11 May 2026 14:39:08 GMT

Anthropic has shipped a lot in the last twelve months. Skills, MCPs, Claude Code, Claude Design, Cowork, integrations everywhere.

Easy to forget that none of it works without the layer underneath: the right prompt.

The prompt is still the thing. The model is smarter, the surface area is bigger, but the difference between an AI feature that ships and one that flakes in production is still written in plain English at the top of an API call.

Below are the lessons from a recent Anthropic prompting workshop run by their Applied AI team. Ten components, in a specific order, with the exact prompt examples they showed on stage from a real Swedish car insurance use case.

Inside you will find:

Why Every V1 Prompt Fails
The 10 Layers Inside Every Production Prompt
How to Brief Claude to Avoid Hallucinations
How to Write Instructions That Survive Edge Cases
What to Repeat Right Before Claude Acts
The Three Output-Control Tricks Most Founders Skip

PS: I’m partnering with Deel to fund up to 10 founders with $1,000,000 🤑

Apply now

👋 Just a quick note before you continue!

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

Here, few examples:

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

Let’s begin!

Stress Test Your Startup with Claude

Guillermo Flor — Fri, 08 May 2026 15:16:16 GMT

Anthropic just released 10 agent templates for financial professionals: pitchbooks, IC memos, due-diligence checklists, DCF and comp models, valuation reviewers.

These are the templates an analyst pulls up when their fund is evaluating you.

So run (some of) them on yourself first. Every hole a VC/ PE analyst would flag in your DD is a hole you can find this weekend. Every weak section in their IC memo is one you can rewrite before a partner ever opens the file.

📚 By the end of this guide you’ll have a self-DD operating system: a Claude project that audits your data room like a PE analyst, builds the model the VC will build, generates the IC memo a partner would write about you and simulates the 30 hardest questions you’ll get in the partner meeting.

A note on stage. The qualitative sections (DD checklist, market map, IC memo, mock meeting) run at any stage, while the numerical sections (unit economics, model audit) get sharper with traction.

Inside you will find:

How to Set Up the Self-DD Project
How to Run a PE-Grade Diligence Checklist on Yourself
How to Audit Your Unit Economics Like a PE Analyst
How to Build the Model the VC Will Build
How to Map Your Market Like a Sell-Side Analyst
How to Generate the IC Memo Before They Do
How to Run the Mock Partner Meeting

PS: I’m partnering with Deel to fund up to 10 founders with $1,000,000 🤑

Apply now

👋 Just a quick note before you continue!

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

Here, few examples:

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

Let’s begin!

1. How to Set Up the Project

The 10 Claude Agents to Go from 0 to 1

Guillermo Flor — Wed, 06 May 2026 18:54:26 GMT

We’ve spent the last few months publishing playbook after playbook, one for almost every job a founder ends up doing twice: ICP, pitch, pricing, outbound, content, hiring, support, legal.

Each one stands on its own. Each one solves a real problem.

But the real switch it’s about stopping to use them standalone and starting to wire them together as a system. When the agents are connected in the right order, the output of one becoming the input of the next, a single founder can do in thirty days what used to require a team and six months.

The piece almost everyone is missing is that sequence is what turns 10 templates into one operating system. So this newsletter is the wrap-up.

Ten agents, four weeks, thirty days, from idea to paying customer.

Inside you will find:

Day 0: The Setup
Week 1. Validate the Idea, Listen to the Buyer, Map the Field
1. Agent 1: The Idea Validation Agent (Days 1-3)
2. Agent 2: The Customer Research Agent (Days 4-5)
3. Agent 3: The Competitive Intelligence Agent (Days 6-7)
Week 2. Pitch, Design, Price the Wedge
1. Agent 4: The Pitch Deck Agent (Days 8-10)
2. Agent 5: The Landing Page Agent (Days 11-13)
3. Agent 6: The Pricing Agent (Day 14)
Week 3. The First Customers Without a Sales Team
1. Agent 7: The Outbound Signal Agent (Days 15-18)
2. Agent 8: The Reddit Engagement Agent (Days 19-21)
Week 4. Distribution That Compounds
1. Agent 9: The SEO + GEO (Days 22-26)
2. Agent 10: The Content + Virality Agent (Days 27-30)

Launch fast. Design beautifully. Build your company’s website on Framer.

Framer helps teams design, build, and launch their marketing sites lightning fast.

Trusted by companies like Miro, Bilt, and Perplexity.

Speed without chaos: ship pages and updates faster without turning the site into a fragile set of one-off hacks
Reduce dependency: shift routine brand and marketing work out of product engineering queues.
Production-grade foundation: Run real marketing systems (CMS, SEO, performance optimization) with governance and collaboration

👉 Build your company’s site on Framer today

Apply to Claim Your Free Year

👋 Just a quick note before you continue!

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

Let’s begin!

1. Day 0. The Setup

Claude for Meta Ads: The Complete Guide

Guillermo Flor — Tue, 05 May 2026 20:23:20 GMT

Meta just published its official MCP connector.

It plugs directly into Claude, takes sixty seconds to install, and gives full read and write access to your ad account through twenty-nine tools across the entire Marketing API.

The connector by itself is a tool. The connector plus Claude’s skills is a system, and the system can be wrapped around the way you specifically think about your customer, your hooks, your angles, your offer, your brand.

The result is that the work an agency used to charge three to five thousand a month for, the part that takes expertise rather than dashboard navigation, can now be done by a founder with a brief, three skill files, and twenty minutes on a Monday morning.

The rest of this guide is the full workflow, end to end, from the connector install to the weekly cycle that runs itself.

Inside you will find:

Why Meta Ads Just Stopped Being an Agency Job
How to Connect Meta to Claude in 60 Seconds
How to Brief Claude Before Launch
From ICP to Live Campaign: The Founder Workflow End to End
The Three Skills Every Founder Should Teach Claude Before Running Ads
The 30-Minute Monday Routine That Replaces a Media Buyer

A CRM shouldn’t obfuscate your data. It should remove entire layers of busywork. Attio is the AI CRM built to do exactly that.

Connect your email and calendar, and Attio turns every customer interaction into context it can reason over and act on. Whether you’re running product-led growth or enterprise sales, Attio gives you the building blocks to run your ideal GTM system.

Then Ask Attio to:

Analyze sales calls and surface patterns across hundreds of conversations
Prep for any meeting with full context on what was said and what’s at risk
Draft personalized post-call follow-ups in seconds

Ask more from CRM. Ask Attio.

Try Now!

👋 Just a quick note before you continue!

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

Here, few examples:

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

Let’s begin!

1. Why Meta Ads Just Stopped Being an Agency Job

The 8 Lessons From Anthropic’s Claude Code Masterclass (+ Anthropic Pitch Deck)

Guillermo Flor — Mon, 04 May 2026 20:02:19 GMT

Anthropic released a 30-minute masterclass that walks through how they actually use Claude Code internally.

The speaker is Boris Cherny, the engineer who built Claude Code at Anthropic. He uses it daily on the codebase that produces it. Around 80% of Anthropic’s technical staff do the same.

Boris doesn’t pitch the product. He walks through the actual habits behind his daily use. I pulled out the 8 lessons that have real teeth.

And since we’re talking about Anthropic: I’m including their original pitch deck as a bonus. It’s one of the cleanest examples of positioning as strategy in modern tech: how to win trust before you win distribution, turn “safety” into a product moat, and raise institutional capital for a research-heavy business. It’s at the end of this newsletter.

Inside you will find:

Stop Using Claude Code Like Autocomplete. It’s an Agent.
Run These 4 Commands Before You Do Anything Else
The 5 Keybindings Every Daily User Should Know
How to Spend Your First Hour With Claude Code
How to Write Your Weekly Standup in 30 Seconds
How to Get 10x Better Output From the Same Prompt
How to Give Claude Permanent Memory of Your Codebase
How to Make Claude Debug Itself

BONUS: Anthropic Original Pitch Deck

PS: I’m partnering with Deel to fund up to 10 founders with $1,000,000 🤑

Apply now

👋 Just a quick note before you continue!

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

Here, few examples:

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

Let’s begin!

1. Stop Using Claude Code Like Autocomplete (It’s an Agent!)

How to Stop Burning Claude Credits

Guillermo Flor — Sat, 02 May 2026 13:41:05 GMT

In the last few weeks, “Claude is burning through my credits” has become one of those things everyone is saying at once.

One developer tracked his usage and found 98.5% of his tokens were spent re-reading conversation history. Only 1.5% went toward the actual output.

Something did shift on Anthropic’s end. They formalized a peak and off-peak structure that means your five-hour session limits drain faster during weekday mornings. That’s real, and it matters.

But here’s what nobody in those threads is saying: the habits that cause most of the waste were already there before the change. The limits just made the problem visible.

This guide is about the mechanism. Once you understand how Claude actually counts what you send, most of the waste becomes obvious and most of the fixes take less than five minutes.

📚 By the end of this guide you’ll know exactly where your credits are going, which habits to change this week, and how to set up a system that cuts your token spend by half or more without touching your plan or losing any output quality.

Inside you will find:

How Claude Actually Counts What You Send
The 6 Habits That Are Draining Your Usage
How to Cut Your Token Spend Without Changing Your Work
How to Build a System That Saves Credits Automatically
How to Tell If You’re Finally Spending Tokens on What Matters

👋 Just a quick note before you continue!

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

Here, few examples:

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

Let’s begin!

1. How Claude Actually Counts What You Send

The Stripe Wallet Agent (Full Guide)

Guillermo Flor — Thu, 30 Apr 2026 18:41:20 GMT

Yesterday agents got wallets.

Stripe just shipped Link’s wallet for agents.

Agents could already pay before but every implementation was custom, fragile, and siloed. Each product had to rebuild the entire flow from scratch.

Now there’s a shared system, actually built for them. Agents can request, get your approval, and complete a purchase: no workarounds, no custom hacks.

And here’s the best part: Link is already the “Pay” button on millions of sites. Merchants don’t need to add anything. If Link is there, agents can already use it.

📚 Inside the guide: the universal pattern for setting up an agent that can buy (with spending controls inline so it doesn’t burn your money), five Claude Code templates you can scaffold this week, and the new business categories just opened by agentic payments.

Inside you will find:

How Stripe Just Turned Agents Into Buyers
How Link’s Wallet for Agents Works
How To Set Up An Agent That Can Buy
The 5 AI Agent Templates Unlocked
What Founders Can Build Now With Agentic Payments

PS: I’m partnering with Deel to fund up to 10 founders with $1,000,000 🤑

Apply now

👋 Just a quick note before you continue!

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

Here, few examples:

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

Let’s begin!

1. How Stripe Just Turned Agents Into Buyers

The Claude Playbook to Predict Virality

Guillermo Flor — Wed, 29 Apr 2026 16:11:53 GMT

Meta just dropped something wild.

An open-source model that predicts how your brain reacts to content before you publish it.

The model is trained on 1,000 hours of fMRI scans from 720 real volunteers watching movies, listening to podcasts, and reading text, it predicts large-scale patterns of brain activity over time when you consume something.

A creator on X took an already-edited AI UGC video draft, ran it through the model, looked at the engagement curve, moved the highest-engagement moment to the front, cut the dead zones where the brain went flat, and reposted the new cut. The views were insane.

📚 By the end of this guide you’ll have a working TRIBE v2 + Claude system: drop in 10 hook variations and get back a ranked list with the neural score and the rewrite suggestion. The same flow runs across thumbnail mocks, ad copy paragraphs, and video drafts.

Inside you will find:

What TRIBE v2 Actually Is
The 30-Minute Setup
The 4 Claude Projects to Pre-Test Hooks, Thumbnails, Videos and Ad Copy
How to Read the Neural Map (and What to Change When Engagement Drops)
The CLAUDE.md That Ties the Whole Stack Together

👋 Just a quick note before you continue!

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

Here, few examples:

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

Let’s begin!

1. What TRIBE v2 Actually Is

YC Startup Screening with Claude

Guillermo Flor — Tue, 28 Apr 2026 20:40:54 GMT

When you analyze a YC batch, the work falls almost entirely on the founders.

Markets shift. A category that looks small this week opens up the next, and the one everybody chased a year ago has quietly closed in on a single winner you weren’t watching.

Traction is even less stable. A company can have zero customers today and a viral February. A strong revenue chart this quarter can collapse on churn the moment the AI demand curve adjusts and the spike turns out to have been the spike, not the trend.

The only signal that holds across that uncertainty is who built the company.

The tier system in this guide is what makes that signal visible at scale. When you’re scrolling through ~150 startups per batch and 300+ founder profiles, the details that actually predict (the second exit, the prior shutdown that wasn’t a failure, the years spent at the top of competitive gaming, the unusual biographical fact buried in a LinkedIn About section) are the same details a human eye drops first.

The system surfaces them, organizes them into four tiers, and renders the result as a kanban your team can keep open. The judgement call after that is still yours.

📚 By the end of this guide you’ll have a continuous YC batch scout: a Cowork artifact with an agent behind it that re-scrapes the batch directory on every reopen, reads every founder’s profile, applies a four-tier classification you can edit, and renders the result as a live kanban your team can pin and check every morning.

Inside you will find:

How to Set Up Cowork as Your Batch Scout
The Four-Tier Logic for Spotting the Founders Worth Meeting
How to Scrape the YC Directory and Every LinkedIn Profile
How to Turn the Scraped Data Into a Tier-Ranked List
How to Build the Final Dashboard
How to Keep the System Alive Across YC Batches

👋 Just a quick note before you continue!

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

Here, few examples:

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

Let’s begin!

1. How to Set Up Cowork as Your Batch Scout

Claude Agent for HR

Guillermo Flor — Mon, 27 Apr 2026 15:50:05 GMT

Hiring a single role eats 15 to 20 hours of founder time.

Most of that time doesn’t decide quality. Writing the job spec takes two hours. Reading two hundred CVs takes another four. Sending status emails, drafting the take-home, writing rejection feedback, building the interview kit, writing the offer letter, planning the first ninety days.

None of this is the work that actually picks who walks in on day one. The work that does pick (interviews, debriefs, reference calls) gets the smaller share, often the tired share.

The gap is not about effort. The variable that moves it is leverage: which parts of the funnel you keep doing yourself, and which parts you hand to a system that has the right context about your company, the role, and the candidate.

📚 By the end of this guide you’ll have an AI-assisted hiring operating system running inside Claude Code: a project that interviews you to write the job spec, scores CVs against your rubric, drafts every candidate email and offer letter, prepares the interview kit, and writes the onboarding plan from the same scorecard that hired the person.

Inside you will find:

How to Set Up the Hiring Agent
How to Write a Job Description That Filters Before Screening
How to Build a Landing Page for Your Open Position
How to Scan Hundreds of CVs Without Losing Quality
How to Prepare Interviews With Claude
How to Automate Every Email From Application to Offer
How to Onboard From the Same Scorecard That Hired

PS: I’m partnering with Deel to fund up to 10 founders with $1,000,000 🤑

Apply now

👋 Just a quick note before you continue!

🤖 3+ New AI agent Templates Every Week
📚 A Growing Library of 20+ AI Agents

Here, few examples:

We’re already 70,000+ founders and operators inside.

Get 7 day free trial

The rest of the article is only available to premium subscribers.

Let’s begin!

1. How to Set Up the Hiring Agent

There is a structural reason your first ten hires are uneven, and it has very little to do with whether you’re a good judge of character. The reasons are:

time poverty → founders focus on shipping the product and hiring gets the residual hours
poor calibration → pattern matching that takes a recruiter three seconds takes the founder fifteen minutes or more
absence of a system

The leverage move is to fix the third reason. You can’t buy back the founder hours. You can’t manufacture five years of recruiter calibration. What you can build is an AI-assisted hiring operating system.