Project Kickoff Gap Analysis
Preamble
A kickoff gap brief makes me write my first judgment before the model gives advice. It turns a new project from a rush of plausible work into a recorded decision, a missing-context list, and one external probe.
The First Frame Is Mine
A new project is where the shortcut starts early. I can give a model a rough idea, receive a tidy plan, and mistake the plan for understanding. The shape arrives before the judgment. The first tasks look obvious. The build begins with borrowed certainty.
The kickoff brief forces the first frame back into my hands before Coach answers: I have to name the decision I think I am making, the constraint I think matters, the risk I am tempted to ignore, and the signal that would prove the project has moved. That record gives the machine something to challenge. It also gives me something to compare against after the answer arrives.
Four Gaps Before Work Starts
Coach should return four kinds of gap. Known missing context is the thing I already know I have not supplied. Unknown-unknown probes are the questions the project might punish me for failing to ask.
LLM blind spots are places where the model is likely to sound confident because the source trace is thin, stale, private, or too abstract. Decision traps are the human shortcuts: premature scope, false precision, feature-first thinking, taste avoidance, or hiding commercial risk inside technical motion.
Each gap has to point at action. A gap that cannot change the first move is commentary.
The Probe Has To Touch The World
The kickoff brief earns its keep through the first probe. That probe can be a buyer question, a demo, a landing test, a direct ask, a prototype shown to the right person, or a small artifact that creates external response.
Coach can recommend the probe, but I still have to choose it. The useful version makes the next contact with reality sharper. The failed version gives me a cleaner plan for staying inside the project.
The Receipt Before Momentum
Every kickoff should leave a receipt: what I believed before Coach answered, which sources Coach read, which gap mattered most, what I accepted, rejected, modified, or deferred, what probe I chose, and what stayed undecided. That receipt protects the build from retroactive storytelling. If the first week goes nowhere, I should be able to see whether the failure came from weak context, weak judgment, weak signal, or a decision I refused to make.
The First Proof
This page is proven by one real kickoff where Coach changes the first build decision before execution. The proof is small by design: a pre-judgment, selected context, four gap types, one first probe, and a decision log tied to 007. The artifact succeeds when the next move becomes harder to fake.