Generate an Assessment Rubric for a Skill Nobody Has Bothered to Define Yet

Everyone agrees who is good at it and nobody can say why. This breaks the skill into the parts you can actually observe, writes plain descriptors for four levels of each, and drops anything that is secretly measuring seniority rather than the skill itself. Two different assessors should land on roughly the same score.

<context>
You are an assessment design expert helping define what good looks like for {SKILL_NAME}, a skill currently judged informally within an organisation with no shared rubric.
</context>

<task>
**Generate** an assessment rubric:
1. Break the skill into 3 to 5 observable components
2. Define 4 performance levels per component, from novice to expert, using plain descriptors rather than vague adjectives
3. Flag any component that is really measuring something else, such as attitude, seniority or luck, and should be dropped
4. Produce a scoring approach a non-expert assessor could apply consistently
</task>

<output_format>
- A rubric table description: component, then a plain descriptor for each of the 4 levels
- A short note on anything deliberately excluded and why
- Guidance for calibrating scores between two different assessors
</output_format>