diff --git a/.github/scripts/close_duplicate_issues.py b/.github/scripts/close_duplicate_issues.py
index ec522af4f88..fd536e42a9f 100755
--- a/.github/scripts/close_duplicate_issues.py
+++ b/.github/scripts/close_duplicate_issues.py
@@ -40,27 +40,16 @@ def gh(*args: str) -> str:
 
 
 def fetch_open_issues(repo: str | None) -> list[dict]:
-    """Fetch all open issues (excluding PRs) via gh api --paginate."""
+    """Fetch all open issues (excluding PRs) via gh api --paginate --slurp."""
     if repo:
         endpoint = (
             f"repos/{repo}/issues?state=open&per_page=100&sort=created&direction=asc"
         )
     else:
         endpoint = "repos/{owner}/{repo}/issues?state=open&per_page=100&sort=created&direction=asc"
-    cmd = ["api", "--paginate", endpoint]
-
-    raw = gh(*cmd)
-    # gh --paginate concatenates JSON arrays, so we may get multiple arrays
-    issues = []
-    for line in raw.strip().splitlines():
-        line = line.strip()
-        if not line:
-            continue
-        parsed = json.loads(line)
-        if isinstance(parsed, list):
-            issues.extend(parsed)
-        else:
-            issues.append(parsed)
+    raw = gh("api", "--paginate", "--slurp", endpoint)
+    pages = json.loads(raw)
+    issues = [issue for page in pages for issue in page]
 
     # Filter out pull requests (they also appear in the issues endpoint)
     return [i for i in issues if "pull_request" not in i]
diff --git a/tests/test_litellm/test_github_close_duplicate_issues.py b/tests/test_litellm/test_github_close_duplicate_issues.py
new file mode 100644
index 00000000000..892c714449e
--- /dev/null
+++ b/tests/test_litellm/test_github_close_duplicate_issues.py
@@ -0,0 +1,108 @@
+"""Unit tests for `.github/scripts/close_duplicate_issues.py`.
+
+These exercise the pure logic (issue fetching and duplicate detection) without
+hitting GitHub. The `gh` CLI wrapper is stubbed via monkeypatch.
+"""
+
+from __future__ import annotations
+
+import importlib.util
+import json
+import sys
+from pathlib import Path
+
+import pytest
+
+SCRIPT_PATH = (
+    Path(__file__).resolve().parents[2]
+    / ".github"
+    / "scripts"
+    / "close_duplicate_issues.py"
+)
+
+
+@pytest.fixture(scope="module")
+def dup_module():
+    """Load the script as a module via its file path (it lives outside the package)."""
+    spec = importlib.util.spec_from_file_location("close_duplicate_issues", SCRIPT_PATH)
+    assert spec and spec.loader, f"Could not load spec for {SCRIPT_PATH}"
+    module = importlib.util.module_from_spec(spec)
+    sys.modules["close_duplicate_issues"] = module
+    spec.loader.exec_module(module)
+    return module
+
+
+class TestFetchOpenIssues:
+    def test_should_flatten_all_pages_from_slurp_output(self, dup_module, monkeypatch):
+        pages = [
+            [{"number": 1, "title": "a"}, {"number": 2, "title": "b"}],
+            [{"number": 3, "title": "c"}],
+        ]
+        monkeypatch.setattr(dup_module, "gh", lambda *a: json.dumps(pages))
+        issues = dup_module.fetch_open_issues("owner/repo")
+        assert [i["number"] for i in issues] == [1, 2, 3]
+
+    def test_should_exclude_pull_requests(self, dup_module, monkeypatch):
+        pages = [
+            [
+                {"number": 1, "title": "real issue"},
+                {"number": 2, "title": "a PR", "pull_request": {"url": "x"}},
+            ]
+        ]
+        monkeypatch.setattr(dup_module, "gh", lambda *a: json.dumps(pages))
+        issues = dup_module.fetch_open_issues("owner/repo")
+        assert [i["number"] for i in issues] == [1]
+
+    def test_should_pass_slurp_flag_to_gh(self, dup_module, monkeypatch):
+        captured: dict = {}
+
+        def fake_gh(*args):
+            captured["args"] = args
+            return json.dumps([[]])
+
+        monkeypatch.setattr(dup_module, "gh", fake_gh)
+        dup_module.fetch_open_issues("owner/repo")
+        assert "--paginate" in captured["args"]
+        assert "--slurp" in captured["args"]
+
+    @pytest.mark.parametrize("sep", ["\u2028", "\u2029", "\x85"])
+    def test_should_not_break_on_unicode_line_separators_in_body(
+        self, dup_module, monkeypatch, sep
+    ):
+        # Regression for #32639: gh emits the GitHub API JSON with Unicode
+        # line separators (\u2028, \u2029, \x85) left literal inside issue
+        # bodies. The old parser split the payload on str.splitlines(), which
+        # breaks on those code points, fragmenting the otherwise-valid JSON
+        # array and raising "Unterminated string" from json.loads.
+        pages = [
+            [
+                {"number": 1, "title": "keep me", "body": f"line one{sep}line two"},
+                {"number": 2, "title": "keep me too", "body": "plain"},
+            ]
+        ]
+        raw = json.dumps(pages, ensure_ascii=False)
+        assert sep in raw
+        assert len(raw.splitlines()) > 1
+        assert json.loads(raw) == pages
+        monkeypatch.setattr(dup_module, "gh", lambda *a: raw)
+        issues = dup_module.fetch_open_issues("owner/repo")
+        assert [i["number"] for i in issues] == [1, 2]
+        assert issues[0]["body"] == f"line one{sep}line two"
+
+
+class TestFindDuplicate:
+    def test_should_match_titles_above_threshold_ignoring_bug_prefix(self, dup_module):
+        issue = {"number": 10, "title": "[Bug]: proxy crashes on startup"}
+        candidates = [{"number": 3, "title": "Proxy crashes on startup"}]
+        dup = dup_module.find_duplicate(issue, candidates, threshold=0.85)
+        assert dup is not None and dup["number"] == 3
+
+    def test_should_not_match_below_threshold(self, dup_module):
+        issue = {"number": 10, "title": "streaming responses are truncated"}
+        candidates = [{"number": 3, "title": "add support for new provider"}]
+        assert dup_module.find_duplicate(issue, candidates, threshold=0.85) is None
+
+    def test_should_skip_self(self, dup_module):
+        issue = {"number": 10, "title": "same title"}
+        candidates = [{"number": 10, "title": "same title"}]
+        assert dup_module.find_duplicate(issue, candidates, threshold=0.85) is None