spec (gstack)
- Product
Turns a vague intent into a precise, buildable specification, working through the ambiguities in stages until the requirements are concrete enough to implement against. It assumes you have decided WHAT you want and need it made executable — not whether the idea is worth doing, and not a critique of a plan you already drafted.
A situation it fits
Engineers asked for exact UI states, API fields, and edge-case behavior for a new feature — they need the details so they can start coding.
3 scenarios in the bank answer to spec (gstack). The rest are in the game.
Skills it gets confused with
These share a family with spec (gstack), which is another way of saying they are the ones you might reach for by mistake.
- office-hours (gstack)Pressure-tests a raw product idea the way a sharp investor would — probing who it is for, why now, and what would make it fail — before any planning or building begins. Unlike turning an idea into a spec or grading a written plan, this is the earliest step: deciding whether the idea is worth pursuing at all.
- plan-ceo-review (gstack)Reviews an already-written plan from a founder's chair — is the scope right, are the priorities ordered well, is this the most valuable thing to build now. It judges a finished plan on business and product grounds, where an engineering review would judge that same plan on architecture, and idea pressure-testing would question the concept before any plan exists.
- retro (gstack)Runs a look-back over a stretch of finished work — what shipped, what went sideways, what to change next time — turning recent experience into concrete process improvements. It faces backward at completed work, where idea pressure-testing, spec-writing and plan reviews all face forward at work not yet done.
Knowing what spec (gstack) does is the easy half. Telling it apart from the others under time pressure is the game.
Today's sessionspec (gstack) is part of garrytan/gstack. Licence: MIT. The description above was written for this game, not taken from the skill.