feat(recipes): delegate CivitAI-image re-import to companion browser extension

Recipes imported from CivitAI image URLs can contain 0 LoRAs: the backend
only sees the REST image API + EXIF, while the complete generation data
lives in the image page's internal trpc payload (see
docs/recipe-civitai-image-no-metadata.md). When the companion
lm-civitai-extension is installed with a valid license, re-import (single
and bulk) of CivitAI-image-sourced recipes is now delegated to the
extension via DOM CustomEvents; the extension scrapes the image page with
the user's session and calls back into the reimport endpoint with the
full metadata payload. Without the extension (or with an invalid license)
the native path runs unchanged.

- POST /api/lm/recipe/{id}/reimport accepts optional payload params
  (image_url/name/resources/gen_params/base_model/tags); the payload path
  reuses the import-remote engine with reimport semantics (user-edit
  carryover, delete-after-save), and malformed/failed payloads fall back
  to the legacy URL import. Response gains loras_count.
- The endpoint also accepts GET: the extension is GET-only by convention
  (documented in AGENTS.md).
- New static/js/utils/extensionReimportBridge.js (probeExtension /
  delegateReimport / getCivitaiImageInfo) wired into RecipeContextMenu
  and BulkManager with silent native fallback.
- i18n: toast.recipes.reimportingViaExtension added and translated in
  all 9 locales.

tests/routes/test_recipe_routes.py

@@ -2341,3 +2341,147 @@ async def test_get_recipe_detail_includes_recipe_json_path(
         assert response.status == 200
         payload = await response.json()
         assert "recipe_json_path" not in payload
+
+
+async def test_reimport_with_extension_payload_uses_payload_path(
+    monkeypatch, tmp_path: Path
+) -> None:
+    """A re-import carrying the companion extension's metadata payload must
+    use the payload-based import engine (caller-supplied LoRAs) instead of
+    the legacy CivitAI image URL import, and report loras_count."""
+    provider_calls: list[str | int] = []
+
+    class Provider:
+        async def get_model_version_info(self, model_version_id):
+            provider_calls.append(model_version_id)
+            return {}, None
+
+    async def fake_get_default_metadata_provider():
+        return Provider()
+
+    monkeypatch.setattr(
+        "py.recipes.enrichment.get_default_metadata_provider",
+        fake_get_default_metadata_provider,
+    )
+
+    async with recipe_harness(monkeypatch, tmp_path) as harness:
+        old_file = harness.tmp_dir / "recipes" / "sub" / "rec-ext.webp"
+        harness.scanner.recipes["rec-ext"] = {
+            "id": "rec-ext",
+            "title": "Old title",
+            "file_path": str(old_file),
+            "tags": ["tag1"],
+            "source_path": "https://civitai.com/images/12345",
+        }
+        harness.civitai.image_info["12345"] = {
+            "id": 12345,
+            "url": "https://image.civitai.com/x/y/original=true/pic.png",
+            "type": "image",
+        }
+        harness.persistence.save_result = SimpleNamespace(
+            payload={"success": True, "recipe_id": "new-rec-ext"}, status=200
+        )
+        # The freshly saved recipe as the scanner would see it (for loras_count).
+        harness.scanner.recipes["new-rec-ext"] = {
+            "id": "new-rec-ext",
+            "loras": [{"file_name": "Painterly"}],
+        }
+
+        resources = [
+            {
+                "type": "lora",
+                "modelId": 20,
+                "modelVersionId": 44,
+                "modelName": "Painterly",
+                "modelVersionName": "v2",
+                "weight": 0.5,
+            },
+        ]
+        # The extension only issues GET requests (per its API convention).
+        response = await harness.client.get(
+            "/api/lm/recipe/rec-ext/reimport",
+            params={
+                "image_url": "https://civitai.com/images/12345",
+                "name": "Extension Recipe",
+                "resources": json.dumps(resources),
+                "gen_params": json.dumps({"prompt": "from extension"}),
+                "base_model": "Flux",
+            },
+        )
+        payload = await response.json()
+
+        assert response.status == 200
+        assert payload["success"] is True
+        assert payload["old_recipe_id"] == "rec-ext"
+        assert payload["recipe_id"] == "new-rec-ext"
+        assert payload["loras_count"] == 1
+
+        save_call = harness.persistence.save_calls[-1]
+        # Caller-supplied payload data wins: name, LoRAs, gen params.
+        assert save_call["name"] == "Extension Recipe"
+        assert save_call["metadata"]["loras"][0]["file_name"] == "Painterly"
+        assert save_call["metadata"]["loras"][0]["weight"] == 0.5
+        assert save_call["metadata"]["gen_params"]["prompt"] == "from extension"
+        # Reimport semantics: original source_path and folder are preserved.
+        assert save_call["metadata"]["source_path"] == "https://civitai.com/images/12345"
+        assert save_call["target_dir"] == str(harness.tmp_dir / "recipes" / "sub")
+        # The old recipe is deleted and user edits carried over.
+        assert harness.persistence.delete_calls == ["rec-ext"]
+        assert harness.persistence.update_calls[-1]["recipe_id"] == "new-rec-ext"
+        assert harness.persistence.update_calls[-1]["updates"]["title"] == "Old title"
+        assert harness.persistence.update_calls[-1]["updates"]["tags"] == ["tag1"]
+
+
+async def test_reimport_with_malformed_payload_falls_back_to_legacy(
+    monkeypatch, tmp_path: Path
+) -> None:
+    """Malformed resources JSON must be treated as "no payload": the legacy
+    source-URL import runs and the request still succeeds."""
+    async def fake_get_default_metadata_provider():
+        return SimpleNamespace(get_model_version_info=lambda id: ({}, None))
+
+    monkeypatch.setattr(
+        "py.recipes.enrichment.get_default_metadata_provider",
+        fake_get_default_metadata_provider,
+    )
+
+    async with recipe_harness(monkeypatch, tmp_path) as harness:
+        harness.scanner.recipes["rec-bad"] = {
+            "id": "rec-bad",
+            "title": "Broken payload",
+            "file_path": str(harness.tmp_dir / "recipes" / "rec-bad.webp"),
+            "tags": [],
+            "source_path": "https://civitai.com/images/12345",
+        }
+        harness.civitai.image_info["12345"] = {
+            "id": 12345,
+            "url": "https://image.civitai.com/x/y/original=true/pic.png",
+            "type": "image",
+        }
+        harness.persistence.save_result = SimpleNamespace(
+            payload={"success": True, "recipe_id": "legacy-new"}, status=200
+        )
+        harness.scanner.recipes["legacy-new"] = {"id": "legacy-new", "loras": []}
+
+        response = await harness.client.get(
+            "/api/lm/recipe/rec-bad/reimport",
+            params={
+                "image_url": "https://civitai.com/images/12345",
+                "name": "Ignored Name",
+                "resources": "{not valid json",
+            },
+        )
+        payload = await response.json()
+
+        assert response.status == 200
+        assert payload["success"] is True
+        assert payload["recipe_id"] == "legacy-new"
+        assert payload["loras_count"] == 0
+
+        save_call = harness.persistence.save_calls[-1]
+        # Legacy URL path: the payload name is ignored and the title is
+        # derived from the (empty) metadata, and no caller LoRAs are used.
+        assert save_call["name"] == "Civitai Image 12345"
+        assert save_call["metadata"]["loras"] == []
+        assert save_call["metadata"]["source_path"] == "https://civitai.com/images/12345"
+        assert harness.persistence.delete_calls == ["rec-bad"]