In use, every project
Built by me
Idea to launch
My own system for taking a product from a raw idea to launch, so no step gets skipped. It started with the skills I built after Ekam and came together on docandthebone. Now it’s a plugin I start every product with.
FOR
Me first, then the team I work with
MADE WITH
Claude, Claude Code
TIMELINE
Oct 2026, updated after every launch
ROLE
Author, and its main user so far
Why I built it
Ekam is where this started. I skipped steps to get to a prototype faster, then spent iteration after iteration getting the design consistent. So I started building skills for the steps I kept skipping: referencing, benchmarking, idea validation, a few of them honestly out of curiosity. docandthebone was the first product I took through all of them in order, and it went from idea to launch in ten days. After that I wrote the whole thing down as one system.
Who it’s for
Anyone starting a product from zero
Built for web apps and websites first. It asks a few questions, shows the whole path, then runs one phase at a time with an approval gate after each.
Build log
How this actually came together
Early 2026
The first skills
Built around Ekam, one for each step I’d rushed past: referencing, benchmarking, idea validation.
Sept–Oct
Run once, in order
docandthebone was the first product to go through every step, before the system had a name.
Oct
First version
Written up from that build, with the setup lessons folded in.
19 Oct
Referencing, always
Every reference board now runs through my own reference skill.
21 Oct
Strategy before scope
Product strategy and a feature map now come before the PRD.
Every launch
A retro feeds it
Dated lessons go back in, so the next product starts smarter.
Decisions
A few calls I had to make along the way
Decision 01
Approval gates, not autopilot
Every phase ends with me signing off. The system drafts, I decide.
Decision 02
It teaches as it goes
Each step explains itself, so it grows my skills instead of hiding the work.
Decision 03
Depth on a dial
The same path runs as a weekend sprint or a full build. Only the depth changes.
Decision 04
Every launch makes it better
A retro after each launch writes dated lessons back into the system.
Me and AI
It starts in my head, not in a prompt
The system is my way of working, written down: my thinking first, then Claude, then an in-depth review before anything moves to the next step.
Writing it
01
S
Me
Had the steps clear in my head from Ekam’s rush and docandthebone’s build
AI
Claude
Turned my notes into phases and skills
S
Me
Rewrote, cut and reordered until it matched how I actually work
Final callMe
Using it
02
S
Me
Bring the idea, the context and the references, and think it through before the first prompt
AI
Claude
Does the groundwork for each phase: drafts, maps, screens and docs
S
Me
Go through every output in depth, then approve or redirect at each gate
Final callMe
Where it stands
I use it on every new product now
It runs as a Claude plugin I start each product with. The details stay private; I’m happy to walk through it on a call.
What I learned
01
A process is only real once it has survived a launch.
02
The best time to write it down was the day after shipping, while every mistake was still fresh.
Want the
full story?
I'm happy to walk you through the files, the trade-offs and what I'd change today. Or tell me about something you're building.
Message me ↗
Based in
Mumbai, working with teams anywhere