| name | new-estimator |
| description | Scaffold a new ICC estimator with its estimand block, test skeleton, and oracle checklist. Use when adding a new ICC variant (e.g. consistency, one-way, multilevel). |
| allowed-tools | Read, Grep, Glob, Edit, Write |
Instructions
Scaffold a new estimator without leaving it un-oracled (refuse otherwise —
PRINCIPLES.md #1). Produce, in order:
- Estimand block first (PRINCIPLES.md #2, before any code): a short spec
naming the population quantity — which variance components are in the
numerator/denominator, single vs. average, agreement vs. consistency, fixed vs.
random raters, and the identifiability conditions. Follow the pattern in
project/estimand-specs/. Represent the ICC as (signal component, {error
component set}, averaging divisor) so variants are choices of the error set and
divisor, not new code paths.
- Function stub in
R/ with roxygen including the "Which ICC is this, and
when should you use it?" note and the key tradeoff (PRINCIPLES.md #13), cli
messaging and classed errors via the abort_*() layer (PRINCIPLES.md #8).
- Test skeleton in
tests/testthat/ with the required oracle checklist
(≥2 independent types): textbook/analytic, established package, seeded
simulation. The oracle plan lives in the estimand-spec (its oracle-set section);
register a row in project/REFERENCES.md only once the oracle is asserted
green, naming the test file and provenance (ADR-015 single-source — REFERENCES
holds no "planned" state; PRINCIPLES.md #4).
- A
project/DECISIONS.md stub if the estimator involves a modeling choice.
Add the task to the active milestone's board in project/MILESTONES.md (ADR-015;
there is no separate TASKS.md). Do not implement the numerics until the estimand
and oracle plan are written.
