Loop & Gate

loop & gate · grow kit

The Grow Kit.

A judgment layer for taking software to market with an agentic loop. Anyone can install the tools that generate design and draft copy. Almost nobody knows where to intervene — whether the design is actually yours, whether the copy is true, which audience is really yours, and whether this is worth shipping now.

Install it

what it is

the other half of the loop

A growth loop can generate designs, draft posts, spin variants, and schedule them on its own, and it should. But it can't tell tasteful from on-trend slop, it can't feel when a claim is technically true but sounds like a lie, it can't read which audience is actually yours, and left alone it will confidently blast a good message into the wrong channel at the wrong time. So the loop stops at fixed points and waits for you.

This is the build side's mirror image. In Loop & Gate, the Build Kit takes an idea to shipped software; this kit takes shipped software to the right people and reads whether it worked. They close a loop through the Foundation vault: this kit's last gate writes market signals where the Build Kit's first gate picks them up to decide what to build next.

It is not the design or copy tooling — it sits on top of two public skills for that — and it is not an autopilot. Every gate stops for your decision.

it needs the Foundation

not optional

The design gate checks output against your taste profile and the copy gate against your voice profile, and both live in the Foundation. Set that up and run its profile interview first, or those two gates are just guessing. The kit's own tools reference says the same thing twice: Grow requires the Foundation, it doesn't just recommend it.

The Foundation's getting-started page covers the install, the vault, and the profile interview — step 6 is the one this kit depends on.

the gates

three rhythms

Unlike the build side, these don't run in one straight pass. Growth is three cadences, and a gate belongs to the rhythm it runs on. Gate ∞ runs every time and decides how much a given piece or campaign earns: a reply in a thread gets a glance at voice and a send; a launch page gets the full per-piece set.

the grow gates, by cadence
CadenceGateThe questionLens
every time ∞ How much does this piece deserve? Read reach and permanence first, and let them set the process. All, above the others
strategy
(set periodically, reused)		 1 Who's the audience, and what do they need to hear? A person in a situation, mapped with the four switching forces — push, pull, habit, anxiety — and written to the force that is actually binding. Positioning
6-define What would “it worked” even mean? Name the observable outcome before you make the thing, so the feedback cadence has something real to read. Positioning
per-piece
(every artifact)		 0 Ready to show? Is the destination worth the attention you're about to drive to it? A great post pointing at a rough landing page wastes a first impression. Experience
2 What's the one claim? One piece, one point — the one true thing that speaks to the binding force, not the longest feature list. Positioning
3 Does the design carry it, on-brand, not slop? On-brand isn't “looks nice,” it's “could only be us.” Design
4 Is the copy true, in voice, nothing fabricated? Check every specific against reality — a number, a testimonial, a capability claim. Content
5 Right channel, format, time? Put the piece where the audience already gathers, in a window you can be around to reply in. Marketing / Content
feedback
(continuous)		 6-measure What happened, and did it feed the next build? Read the outcome you named at 6-define, and write the signal where the build side will read it. Positioning and Marketing

Order for a single artifact: ∞ → 0 → 2 → 3 → 4 → 5. On a rhythm, run 6-measure and feed what you learn back to the build side. That last step is where the loop closes.

install

Foundation first

Set up the Foundation

Not a prerequisite you can skip: the profiles this kit reads come from there. Follow the Foundation's assembly order, and run the profile interview — a conversation about how you write and how you judge good work — before coming back.

Add the kit

In the Code chat box:

/plugin marketplace add slogsdon/loop-and-gate-grow-kit

Click Install on the menu that appears.

Add the design and copy skills

Both are public, through the slogsdon/claude-code-config marketplace, and both read the profiles you built in step 1. skills-design generates on-brand design; skills-writing drafts and cleans up copy in your voice.

/plugin marketplace add slogsdon/claude-code-config
/plugin install skills-design@slogsdon-claude-code-config
/plugin install skills-writing@slogsdon-claude-code-config

Run it

/grow-and-gate

Tell it whether you're setting strategy, making a piece, or reading results. Strategy is once per campaign; a piece is every artifact; results are a rhythm.

On pi, one command installs the kit and the two skills are separate pi install lines listed in reference/tools.md:

pi install git:github.com/slogsdon/loop-and-gate-grow-kit

a worked example

a launch page

You're about to post a launch page, so you run /grow-and-gate.

Gate ∞. High reach, permanent, hard to walk back — it earns the full per-piece set. Gate 2. The agent lists five things you could say, and you cut to the one claim that speaks to your audience's real hesitation. Gate 3. The design looks clean but generic; checked against your taste profile it's “could be anyone,” so it goes back until it's yours.

Gate 4. The copy is persuasive and contains a statistic you never gave it. Cut the invented number, rewrite in your voice. Gate 5. One channel where your audience actually is, at a time you can be around to reply.

A quick reply, run through the same skill, earns a glance at voice and a send. That difference is Gate ∞ doing its job.

the tool footprint

most gates have no tool

The kit's own snapshot was verified against a live setup on 2026-07-12 and is labelled disposable on purpose. Its honest read-out: audience, the one claim, ready-to-show and how-much-process are pure judgment with no tool at all. Only design and copy have a skill behind them, scheduling has an optional free-tier account, and metrics use whatever is already free.

So the whole footprint is two public plugins, one optional free account, and free platform analytics. You can work every gate with no paid tools, and every gate with no external accounts at all if you post by hand. When the tool names change, the gates don't.

where to go

after the install