The definition, and what it excludes
A skills assessment is a structured test that establishes whether a person can perform a specific skill at a specific level. Three words in that sentence are doing the work. Structured, meaning the same questions and the same standard for everyone. Specific skill, meaning one capability rather than a role or a department. Specific level, meaning it is measured against a written definition of what that level requires.
This rules out several things it is often confused with. A performance review assesses a person's contribution over a period, which is a different question. A self-rating survey collects opinions about capability. A certification tests a body of knowledge, usually against an external syllabus. Each is useful; none of them tells you whether this person can do this thing to this standard right now.
Why the level matters more than the score
A percentage is only meaningful against a threshold, and a threshold is only meaningful against a definition. Seventy percent on an unspecified test of “JavaScript” tells you nothing you can act on, because nobody has said what the test was pitched at.
This is why levels come first. In PDP Quest each skill is defined at four rungs — Beginner, Intermediate, Advanced and Expert — and an assessment is written for one of them. Passing takes 4 correct out of 5, and the questions are pitched at that rung specifically: ask an Expert something a Beginner could reason out and you have learned nothing about either.
What a good assessment question looks like
The craft is mostly in the wrong answers. Three obviously incorrect options make a one-option question, and a person who knows nothing will still pass it. Every distractor should be something a real practitioner would plausibly pick, for a reason you could name out loud.
The fastest way to find out whether a question works is to write the explanations before the question itself. If you cannot say why an option is wrong, it is not a distractor, it is filler — and a test built from filler measures reading speed rather than skill.
What you do with the result
On its own, a verified level is a fact about one person on one day. It becomes useful when it is compared with something: the level the role requires, the level the person held last quarter, or the levels held across a team.
That comparison is the whole point. An assessment that produces a number nobody compares to anything has cost somebody an afternoon and changed nothing. The next step after measuring is always a plan against the gap it revealed, with a target level and a date.