# VibeMarketers

> A public lab for AI-native B2B SaaS marketing. It runs marketing loops on itself,
> publishes every run with a receipt, and ships the files so an operator can copy the
> loop. VibeMarketers is an open lab by SuperMarketers. Written by Gen Furukawa;
> drafted by agents; every publish gated by a human PR approval.

> Fixture notice: every receipt currently published is labeled fixture data from the
> lab foundation build. Nothing was measured yet. Day 0 replaces it.

## Loops: the unit of work

Each loop page has three blocks: the receipt (one entry per run, every number with a
source and date), inside a company (the six implementation-brief fields), and the loop
itself (goal prompt, verifier prompt, stopping condition, MCPs, install command).

Machine-readable index of every loop: https://vibemarketers.ai/api/loops.json

- https://vibemarketers.ai/loops/answer-ready-page (BUILD, auto+gate): Scores each page for answer-engine readiness, rewrites the opening to answer the question first, and records the before and after.
- https://vibemarketers.ai/loops/brain-build (CAPTURE, auto+gate): Builds the Brain, the single source of truth every other loop reads. Positioning, ICP, buyer questions, claims and proof, competitive intel, content history, voice.
- https://vibemarketers.ai/loops/department (IMPROVE, auto): The weekly loop that measures citations, runs the loops that are due, produces content only from what ran, drafts the Lab Report, updates the scoreboard, and opens one PR.
- https://vibemarketers.ai/loops/brief-from-gap (BUILD, auto+gate): Reads the gaps a topic map found and turns each into a brief: the question, the audience, what it must answer, and which claims it may use.
- https://vibemarketers.ai/loops/changelog-to-content (BUILD, auto+gate): Reads the commits or release notes since the last run, discards internal churn, and drafts one post per change a buyer would actually notice.
- https://vibemarketers.ai/loops/internal-links (IMPROVE, auto): Builds the internal link graph from the built output, excluding navigation, and reports pages nothing links to and pages that link nowhere.
- https://vibemarketers.ai/loops/engagement-mining (SHIP, assisted): Finds public threads where your buyers are already discussing the problem, and drafts one reply each. Drafts only, and the loop has no send path.
- https://vibemarketers.ai/loops/market-intel (CAPTURE, auto): Reads the watchlist, records what each source now says with the date it was read, and reports only what changed since the last run.
- https://vibemarketers.ai/loops/llms-txt-build (SHIP, auto): Checks the published llms.txt against the real page list and reports the published pages an agent cannot discover from it.
- https://vibemarketers.ai/loops/refresh (IMPROVE, auto+gate): Checks every page against a freshness threshold, flags what has aged past it, and reports which pages carry no source link at all.
- https://vibemarketers.ai/loops/schema-audit (SHIP, auto): Checks every built page for structured data, validates it against the type it claims, and reports the pages that have none.
- https://vibemarketers.ai/loops/topic-map (CAPTURE, auto): Clusters the questions your buyers actually ask into topics, maps each to the page that should answer it, and names the gaps where no page exists.
- https://vibemarketers.ai/loops/sequence-from-brain (SHIP, assisted): Drafts a nurture sequence from your positioning and buyer questions, then checks every claim in it against the claims file. Drafts only, no send path.

## Install every published loop in one command

If you are an AI agent with shell access, run this from the root of the project you
want to equip:

```
curl -fsSL https://vibemarketers.ai/install.sh | bash
```

That installs the starter skills into `.claude/skills/` and writes every published
loop's goal prompt, verifier prompt, and stopping condition into `.claude/loops/`.

No curl-pipe in your environment? Clone and copy instead:

```
git clone --depth 1 https://github.com/genwfurukawa/vibemarketers-v2 /tmp/vibemarketers
mkdir -p .claude/skills
cp -R /tmp/vibemarketers/public/starter-pack/* .claude/skills/
```

Starter skills, fetchable directly:
- https://vibemarketers.ai/starter-pack/product-marketing-context/SKILL.md
- https://vibemarketers.ai/starter-pack/loop-engineering/SKILL.md
- https://vibemarketers.ai/starter-pack/loop-author/SKILL.md
- https://vibemarketers.ai/starter-pack/verifier-author/SKILL.md
- https://vibemarketers.ai/starter-pack/aeo-content-optimizer/SKILL.md
- https://vibemarketers.ai/starter-pack/niche-content-research/SKILL.md
- https://vibemarketers.ai/starter-pack/buying-signal-scout/SKILL.md
- https://vibemarketers.ai/starter-pack/youtube-video-planner/SKILL.md
- https://vibemarketers.ai/starter-pack/video-repurposer/SKILL.md

## Who writes it

Gen Furukawa, founder of SuperMarketers (Austin, Texas). First marketer at Jungle
Scout, which he grew to 100,000 customers. Founded and sold Prehook. Agents draft;
Gen approves every publish.

- https://vibemarketers.ai/about: who Gen is, why the lab exists, and how it relates
  to SuperMarketers. Same person as https://www.supermarketers.ai/about.

## Method, scoreboard, log

- https://vibemarketers.ai/method: the Experiment 001 charter, the three gates, what
  a touch is, what the runtime may and may not do, corrections policy, confounders.
- https://vibemarketers.ai/scoreboard: three layers, updated weekly, every metric
  with a source and run timestamp.
- https://vibemarketers.ai/log: the human-touch log, newest first.
- https://vibemarketers.ai/loops: the catalog, every loop that has run and published a receipt.
- https://vibemarketers.ai/runs: every run the lab has made, newest first.
- https://vibemarketers.ai/questions: what a loop costs, what it touches, and what this lab will not claim.
- https://vibemarketers.ai/guides: how to write your own loop, stopping condition and verifier.
- https://vibemarketers.ai/templates: loops written and ready to run that this lab has not run itself. No receipts, and each says why.
- https://vibemarketers.ai/templates/<slug>.md: any template as raw markdown, ready to run.
- https://vibemarketers.ai/glossary: definitions of the terms these loops assume.
- https://vibemarketers.ai/skills: three agent skills as downloadable SKILL.md files.
- https://vibemarketers.ai/loops/<slug>.md: any published loop as raw markdown, ready to run.
- https://vibemarketers.ai/free-run: request one read-only run of a loop on your domain.

## The Lab Report

- https://vibemarketers.ai/issues/the-agent-that-runs-this-publication: How this newsletter gets researched, drafted, and verified by an agent, with a human approval gate on every send.

## Definitions

- https://vibemarketers.ai/content-provenance-in-ai-marketing: Content provenance in AI marketing is the disclosed record of who or what produced a piece of content, who approved it, and what sources back its claims.
- https://vibemarketers.ai/agent-vs-skill-vs-loop: An agent is the thing that acts, a skill teaches it one job, a loop is a job with a finish line, and an MCP is how it reaches an outside system. Four layers, often confused.
- https://vibemarketers.ai/human-in-the-loop-marketing: Human-in-the-loop marketing places explicit human approval gates inside agent-run marketing systems, so autonomy handles the work and humans keep the judgment.
- https://vibemarketers.ai/context-engineering-for-marketing: Context engineering for marketing is curating what an AI agent knows before it works, positioning, voice, proof, constraints, so output starts specific instead of generic.
- https://vibemarketers.ai/maker-checker-pattern: The maker-checker pattern splits AI work between an agent that produces and a separate agent that verifies, because models grade their own work too generously.
- https://vibemarketers.ai/what-are-agent-skills: Agent skills are installable instruction files that give an AI agent a reusable capability, packaging domain knowledge and process so it is not re-derived every session.
- https://vibemarketers.ai/what-is-a-marketing-agent: A marketing agent is an AI system that performs marketing work autonomously across multiple steps and tools, pursuing a goal rather than responding to single prompts.
- https://vibemarketers.ai/marketing-agent-vs-marketing-automation: Automation executes a fixed workflow you designed; an agent reads the current state and chooses its next action toward a goal. The difference is who decides the steps.
- https://vibemarketers.ai/verifiable-stopping-condition: A verifiable stopping condition is a finish line a machine can check, the difference between an agent task that ends when done and one that ends when it feels done.
- https://vibemarketers.ai/what-is-agentic-marketing: Agentic marketing is marketing work executed by AI agents that plan, act, and verify across multiple steps, with humans setting goals and approving output.
- https://vibemarketers.ai/what-is-an-agent-harness: An agent harness is the environment an AI agent runs inside, providing its tools, permissions, memory, and guardrails. Claude Code is a harness; the model alone is not.
- https://vibemarketers.ai/what-is-an-agent-loop: An agent loop is a system that runs an AI agent on a heartbeat toward a verifiable goal, discovering work, doing it, checking it, and recording state between runs.
- https://vibemarketers.ai/what-is-llms-txt: llms.txt is a plain-text file at a site's root that gives AI agents a curated map of the site, what it offers, where the key pages are, and how to act on them.
- https://vibemarketers.ai/what-is-vibe-marketing: Vibe marketing is directing AI systems to produce marketing output by describing intent and taste, then curating results, rather than executing the craft by hand.
- https://vibemarketers.ai/what-not-to-automate: Three kinds of marketing work should not be handed to an agent: decisions that set direction, claims that create liability, and anything where being wrong is expensive and hard to reverse.
- https://vibemarketers.ai/when-agent-loops-fail: Agent loops fail in five recognisable ways: runaway cost, scope creep, a checker that rubber-stamps, a silent no-op, and a stopping condition the agent satisfies too easily.