When you fix behavior (especially edge/boundary cases), add regression tests that are: - **Reachable** (not unexported/inaccessible without refactors that make them testable). - **Meaningful** (tests should fail when the buggy logic is restored, not merely “exercise” code). - **Aligned with the repo’s runner** (use the framework actually used by the unit test script). - **Supported by testable design** (prefer dependency injection over `as any` to control external effects). Apply like this: - If you add/adjust internal logic, make it testable (e.g., extract helpers or inject dependencies) so unit tests can target the exact behavior. - For boundary bugs, include exact-limit cases and verify tests catch regressions. Example (boundary regression + meaningful failure): ```ts // Arrange const LIMIT = 10; // Restoring a buggy form should make this test fail. function fixedAppend(current: string, next: string) { if (!next) return current; if (current.length >= LIMIT) { const keep = LIMIT > next.length ? LIMIT - next.length : 0; if (keep <= 0) return next.slice(-LIMIT); return current.slice(-keep) + next; } return current + next; } // Assert (exact boundary) import test from "node:test"; import assert from "node:assert/strict"; test("appendBoundedText clamps at limit", () => { const cur = "x".repeat(LIMIT); const next = "y".repeat(LIMIT); assert.equal(fixedAppend(cur, next), next); // should be clamped }); ``` Also add guard assertions when you have rule systems/transform pipelines so that “no-op” rules (e.g., patterns that never change anything) can’t be introduced unnoticed.