Your agent will make a hundred product decisions today. It won't mention most of them. Plenty of those were yours to make.
These skills make it ask you first, show you the calls it made on its own, and hand the ones that matter back. They're what I use every day, refined on real work.
For PMs, engineers, founders and designers: anyone whose agent builds faster than they can decide.
Every gap you leave, your agent fills. Quietly.
Ask for something big and your agent starts immediately. No questions, no brief, no idea how you'll judge the result. Every gap it hits it fills with a guess, and a guess about what to build is a product decision. You find out which ones it got wrong once the work is built on top of them.
A good product person does the opposite. They ask first. They come back with a goal you can correct. They say plainly which calls are theirs and which are yours. That's all these skills do: they make the agent work that way, so the decisions that matter stay with the person who owns them.
AgentBefore I start: I'm aiming at more visitors starting a trial. I'm assuming the pricing page stays as it is. One thing I need from you — which decisions come back to you?
YouJust pricing. You own the rest.
AgentThen here's the goal I'd work to. Correct anything before I start.
Three are ready to install. Two are in the workshop.
Each one takes a job a product person does by hand and teaches it to your agent. They work in Claude Code, and they're written in plain English: no code required to use them, or to read them.
-
1.
Available now
Context is Queen
Your agent decides what “done” means the moment you stop talking. This one asks instead. It shows you what it's aiming at before it touches anything, interviews you a question at a time, proposes a goal for you to correct, and lists every call it made without you when you ask.
Read the docs → -
2.
Available now
Turning Research Into Strategy
You have call transcripts, a painpoint tracker and a deck claiming eleven customers asked for the thing. This gives up the easy assertion. It counts the evidence under a rule you set first, opens the product and tries it, and traces every number back to the person who said it.
Read the docs → -
3.
Available now
Read The Room
Every app looks the same, because everyone is looking at each other instead of at people. This goes and reads how real people are actually talking right now, in their own words and dated, and hands back the decision that puts in front of you. It never searches your category's keywords, because those belong to whoever is selling into it.
Read the docs → -
4.
In the workshop
Best Use of My Time
Your agent asks you things it never checked, and you end up working out what it meant instead of making the call. This makes it do that work first: verify the premise in the real product, name the exact screen rather than "the onboarding", and lead with its own recommendation — so what reaches you is a decision, not a puzzle.
Read the docs → -
5.
In the workshop
AI Output Clinic
Generic AI copy is what writing sounds like when it treats the reader as a target. Everyone's instinct is to rewrite the prompt. This finds where the failure actually lives first, in the data, the instruction, or the screen it lands on, then pins it as a check so it cannot come back.
Read the docs →
Context is Queen, in about a minute.
Using the Claude app? Download the skill as a ZIP from the latest release, then go to Customize → Skills, click +, choose Create skill, and select Upload a skill. Nothing to install on your machine.
In Claude Code, paste this into the chat instead. “Globally” makes it available in every project, not just this one.
› Install the Context is Queen skill globally from github.com/hgambas/think-like-a-pm
Prefer the plugin system? Run
/plugin marketplace add hgambas/think-like-a-pm, then install
context-is-queen from the /plugin menu. That version
also sets up the session-start nudge, so fresh sessions offer the briefing
on their own.
Installs don't carry across. Adding it in the app doesn't add it to Claude Code, or the other way round.
- 1.Ask for something complex, in your own words
“Rework our onboarding so people actually finish setup.” No special syntax needed. The briefing starts on its own, and simple asks skip it.
- 2.It tells you what it's aiming at
Before it does anything: what it thinks done looks like, what it's assuming, and what it needs from you.
- 3.Answer the questions, one at a time
Three to eight of them, covering only what your agent can't research for itself.
- 4.Correct the goal it proposes
Then it works to that goal, and checks itself against it before claiming anything is done.
“Work aimed at a goal you've corrected is simply better work — and you stay the decision-maker on your own project.”
I'm Hiba Ganta. I'm the product person and the designer; my agent is my engineer. I don't read code, which means I can't catch a wrong turn by reading the diff. I have to catch it in how the work gets decided.
Every skill here started as a rule I kept repeating to my own agent until it was worth writing down properly.