react-doctor/jotai-derived-atom-returns-fresh-object

Split the derivation into per-field primitive derived atoms, or wrap with selectAtom(source, fn, shallow) from jotai/utils when a wrapper object is required.

Validation prompt

Confirm the detector match and collect the required evidence before deciding whether an edit is warranted.

Fires only on a CallExpression to jotai's imported atom (verified via the import binding, local-name aliasing allowed) whose FIRST argument is a function with EXACTLY one Identifier parameter (the get param), where that body both (a) references the get parameter as a call somewhere not nested in an inner function, and (b) returns a fresh structure: a bare ObjectExpression/ArrayExpression literal (including spread { ...get(x) }), or a method chain whose OUTERMOST call is an array-producing instance method (filter, map, flatMap, slice, concat, flat, toSorted, toReversed, toSpliced, with, sort, reverse: note sort/reverse mutate-and-return their receiver rather than allocate) or a static Object.{keys,values,entries,fromEntries,assign,create} / Array.{from,of} call. For block bodies every reachable top-level return must be fresh. The detector never inspects the atom() call's second argument. False positives a reviewer would suppress: a reader function that takes more than one parameter (e.g. (get, set) => ...) or zero params, skipped because only single-Identifier-param readers are inspected: this is why the read-write fixture is also quiet, since its reader (get) => get(baseAtom) additionally returns a reference-stable get(x) member value; chains whose outermost terminator returns a primitive or stable value (.filter(...).reduce(...), .find(), .some(), .includes(), .join(), .items.length); a plain get(x).user member access (reference-stable); a constant atom whose body never calls get; an atom imported from a non-jotai module; and block bodies that mix one fresh return with one reference-stable return.

Evidence boundary

The diagnostic proves only that the detector’s modeled source pattern matched. It does not prove runtime impact, product intent, rendered failure, or that one remediation is correct.

Establish the environment, repository policy, exceptions, and required rendered or runtime evidence before deciding the occurrence.

Record one outcome:

• Confirmed failure: The required evidence establishes the violation.
• Rejected: A documented exception or false-positive predicate applies.
• Needs evidence: Named evidence can still be collected.
• Unavailable: Required evidence cannot be collected in this run.
• Waived with evidence: An authorized, scoped exception applies to an established failure.
• Observation: The review records an optional tradeoff without claiming a defect.

A waiver records its scope, authority, evidence, and review condition. It is not a pass or false positive.

Default severity is registry metadata. Use the occurrence’s JSON severity after repository configuration when ordering real findings.

Fix prompt

Apply this candidate correction only after the required evidence confirms the risk.

jotai propagates derived values with Object.is and has no shallow compare on a plain atom, so a fresh object/array every read re-renders every consumer on every upstream notify even when fields are unchanged. Either split the wrapper into one primitive derived atom per field (e.g. const countAtom = atom((get) => get(cartAtom).items.length) and a separate totalAtom), each of which dedupes via Object.is, or if a single wrapper object is genuinely required wrap it with selectAtom(sourceAtom, (s) => ({ ... }), shallow) from jotai/utils (importing shallow) so equal shapes are deduped. For array chains, memoize the source or move the allocation to the consumer.

Repository-wide copy prompt

Use this repository-wide prompt only after validating each occurrence. For one occurrence, use the guidance above.