name	importer-regression
description	Use when changing importer mapping, synchronizer, merger, ExternalReference matching, deleteExtra or AllowDeletions behavior, deduplication, or re-import/idempotency semantics. Adds the right regression tests and validates importer stability.

Importer Regression Skill

Use this skill when a change touches importer behavior that can affect existing models after re-import.

Goals

Preserve stable behavior for repeated imports.
Protect ID stability and existing model matches.
Catch regressions in matching precedence, deduplication, and deletion paths.

Workflow

Identify the true matching boundary.

Determine whether the change lives in task validation, source parsing, persistable generation, shared synchronization, or importer-specific merge logic.
Find the existing matching key or precedence path, especially ExternalReference, schema-aware lookup, or manual-association preservation.

Inspect the nearest regression tests first.

Look for existing tests that already cover idempotency, duplicate handling, or deletion behavior in the touched importer.
Reuse existing TestData factories and scenario helpers.

Make the smallest behavior change that fixes the root cause.

Prefer changing the actual matching or merge rule rather than adding post-processing cleanups.
Avoid broad refactors unless they are necessary to preserve correctness.

Add targeted regression coverage.

Include at least one focused test for the changed scenario.
If structure changes materially, add or update a snapshot test.
For merge changes, consider both first import and re-import behavior.

Validate like an importer maintainer.

Run the nearest importer tests.
If the change is broad, run a focused cross-section across affected importer suites.

Checklist

Does a re-import preserve IDs where it should?
Are ExternalReference values still stable and unique?
Does deletion behavior still differ correctly between enabled and disabled modes?
Are manually modeled items preserved when tests expect them to be?
Did you add a regression test in the closest existing test suite?

Importer Regression Skill

Use this skill when a change touches importer behavior that can affect existing models after re-import.

Goals

Preserve stable behavior for repeated imports.

Protect ID stability and existing model matches.

Catch regressions in matching precedence, deduplication, and deletion paths.

Workflow

Identify the true matching boundary.

Determine whether the change lives in task validation, source parsing, persistable generation, shared synchronization, or importer-specific merge logic.

Find the existing matching key or precedence path, especially ExternalReference, schema-aware lookup, or manual-association preservation.

Inspect the nearest regression tests first.

Look for existing tests that already cover idempotency, duplicate handling, or deletion behavior in the touched importer.

Reuse existing TestData factories and scenario helpers.

Make the smallest behavior change that fixes the root cause.

Prefer changing the actual matching or merge rule rather than adding post-processing cleanups.

Avoid broad refactors unless they are necessary to preserve correctness.

Add targeted regression coverage.

Include at least one focused test for the changed scenario.

If structure changes materially, add or update a snapshot test.

For merge changes, consider both first import and re-import behavior.

Validate like an importer maintainer.

Run the nearest importer tests.

If the change is broad, run a focused cross-section across affected importer suites.

Checklist

Does a re-import preserve IDs where it should?

Are ExternalReference values still stable and unique?

Does deletion behavior still differ correctly between enabled and disabled modes?

Are manually modeled items preserved when tests expect them to be?

Did you add a regression test in the closest existing test suite?

importer-regression