Pre-existing issue (surfaced during #2026 review)
PREnrichmentData and BatchObserver are each declared twice in the core types. TypeScript interface merging makes this compile, but the duplication is confusing and the first copy of each is an older, less-complete version.
Current location (after the #2026 types split): packages/core/src/types/scm.ts
PREnrichmentData — lines ~126 (older) and ~330 (current, includes ciChecks)BatchObserver — lines ~155 (older) and ~361 (current)
Origin: this is NOT introduced by #2026 — the duplicates already existed in the monolithic types.ts on main (lines 906/1110 and 935/1141). The split relocated them verbatim, which is why they're now more visible side-by-side in scm.ts.
Proposed fix
Delete the older/incomplete first declaration of each interface, keeping the complete version. Verify the merged shape is identical to today's (so no behavior change) — i.e. the older copy is a strict subset of the current one.
Why a separate issue
#2026 is a pure, verbatim relocation with a guaranteed-unchanged public surface; deleting declarations there would break that invariant and the byte-for-byte review guarantee. This cleanup is a small, independent follow-up.
Pre-existing issue (surfaced during #2026 review)
PREnrichmentDataandBatchObserverare each declared twice in the core types. TypeScript interface merging makes this compile, but the duplication is confusing and the first copy of each is an older, less-complete version.Current location (after the #2026 types split):
packages/core/src/types/scm.tsPREnrichmentData— lines ~126 (older) and ~330 (current, includesciChecks)BatchObserver— lines ~155 (older) and ~361 (current)Origin: this is NOT introduced by #2026 — the duplicates already existed in the monolithic
types.tsonmain(lines 906/1110 and 935/1141). The split relocated them verbatim, which is why they're now more visible side-by-side inscm.ts.Proposed fix
Delete the older/incomplete first declaration of each interface, keeping the complete version. Verify the merged shape is identical to today's (so no behavior change) — i.e. the older copy is a strict subset of the current one.
Why a separate issue
#2026 is a pure, verbatim relocation with a guaranteed-unchanged public surface; deleting declarations there would break that invariant and the byte-for-byte review guarantee. This cleanup is a small, independent follow-up.