[dinwwwh]

🤖 Generated with Claude Code

https://claude.ai/code/session_01VVaw7RYWCsn8myBE4Rncuq

Test-only follow-up to #374.

Stryker on #374's isInlined lines (v3.1-to-v3.0.ts:298-325) left 4 survivors. One is the known equivalent name === 0 -> false, unchanged since #370. The other three are test gaps:

• '#/components/schemas/' -> '': the renamed-key check would then apply to every component type, so a schema $ref to #/components/parameters/Query/schema (the key schema is no Schema keyword) would be inlined. No test kept it as written.
• next === -1 -> false, and -1 -> +1: when the ref ends at the key, the key loses its last letter, so #/components/schemas/Negated/not reads no, counts as renamed and is inlined. No test held a $ref that ends at a key 3.0 keeps.

This adds both cases to the "replaces a $ref into a key that becomes an extension" test. Each must stay a bare $ref (toStrictEqual).

• pnpm exec vitest run packages/downgrader: 311 passed
• Stryker on the same lines with this test: 68 of 69 killed (was 65); the one left is the known equivalent
• pnpm run check passes

Separately, refreshing the 72 fuzz dumps on defcf13 changed 0 outputs.

---

Generated by Claude Code

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[codspeed]

Merging this PR will not alter performance

✅ 29 untouched benchmarks
⏩ 2 skipped benchmarks¹

---

_{Comparing claude/project-thread-rixcs1 (9928840) with main (defcf13)}

Footnotes

1. 2 benchmarks were skipped, so the baseline results were used instead. If they were deleted from the codebase, click here and archive them to remove them from the performance reports. ↩

[pullfrog]

ℹ️ No critical issues — the mutation gaps this PR targets are genuinely closed. One rough edge on how strongly the new cases hold.

Reviewed changes

• $ref ending at a kept key: Negated: { not: … } plus IntoKept: ref('Negated/not'), asserted with toStrictEqual to stay a bare $ref.
• $ref into another component type: IntoParameter: { $ref: '#/components/parameters/Query/schema' } with components.parameters.Query added to the input, likewise asserted to stay as written.

Verified locally against packages/downgrader/src/v3.1-to-v3.0.ts: with '#/components/schemas/' → '', with next === -1 → false, and with next === -1 → next === 1, the extended test fails in each case, and unmutated it passes (140 tests, lint clean). That matches the PR description.

  ^{｜ Fix all ➔ ｜ Fix 👍s ➔ ｜ View workflow run ｜ Using deepseek-v4.1-flash-1017 (free via Pullfrog for OSS) ｜ 𝕏}

[pullfrog]

✅ No new issues found.

Reviewed changes

• Kept $ref targets asserted: added Negated to the schema toMatchObject and asserted components.parameters.Query converts with its schema, so neither kept $ref can pass just because its target dangles. This resolves the only open finding from the prior review.
• No source changes: the delta is test-only; isInlined is untouched.

Verified the full v3.1-to-v3.0.test.ts file passes (140 tests) at 9928840.

  ^{｜ View workflow run ｜ Using deepseek-v4.1-flash-1017 (free via Pullfrog for OSS) ｜ 𝕏}