security: keep indexed files inside the cloned repository

.github/workflows/ci.yml

@@ -26,6 +26,10 @@ jobs:
       - name: Compile backend
         run: python -m compileall -q backend
 
+      - name: Run backend unit tests
+        working-directory: backend
+        run: python -m unittest discover -s tests -v
+
   frontend:
     runs-on: ubuntu-latest
     steps:

backend/services/github.py

@@ -73,16 +73,30 @@ def extract_zip(upload_path: Path):
         raise
     return repo_id, dest
 
+def _resolve_repo_file(root: Path, candidate: Path):
+    root = root.resolve()
+    try:
+        resolved = candidate.resolve()
+        resolved.relative_to(root)
+    except (OSError, ValueError):
+        return None
+    if not resolved.is_file():
+        return None
+    return resolved
+
+
 def iter_files(root: Path):
+    root = root.resolve()
     for p in root.rglob("*"):
-        if not p.is_file(): continue
         rel = p.relative_to(root)
         if any(part in SKIP_DIRS for part in rel.parts): continue
-        try: size = p.stat().st_size
+        resolved = _resolve_repo_file(root, p)
+        if resolved is None: continue
+        try: size = resolved.stat().st_size
         except OSError: continue
         if size > 300_000: continue
         if p.suffix.lower() in TEXT_EXTENSIONS or p.name.lower() in {"dockerfile","makefile"}:
-            yield p
+            yield resolved
 
 def read_text(path: Path): return path.read_text(encoding="utf-8", errors="replace")
 

backend/tests/test_repository_boundary.py

@@ -0,0 +1,39 @@
+import tempfile
+import unittest
+from pathlib import Path
+
+from services.github import iter_files
+
+
+class RepositoryFileBoundaryTests(unittest.TestCase):
+    def test_external_symlink_is_not_indexed(self):
+        with tempfile.TemporaryDirectory() as temp_dir:
+            root = Path(temp_dir)
+            repo = root / "repo"
+            repo.mkdir()
+            outside = root / "outside.py"
+            outside.write_text("SECRET = True\n", encoding="utf-8")
+            link = repo / "linked.py"
+            try:
+                link.symlink_to(outside)
+            except (OSError, NotImplementedError):
+                self.skipTest("symlinks are unavailable on this platform")
+
+            indexed = list(iter_files(repo))
+
+            self.assertEqual(indexed, [])
+
+    def test_internal_file_is_still_indexed(self):
+        with tempfile.TemporaryDirectory() as temp_dir:
+            repo = Path(temp_dir)
+            source = repo / "src" / "main.py"
+            source.parent.mkdir()
+            source.write_text("print('ok')\n", encoding="utf-8")
+
+            indexed = list(iter_files(repo))
+
+            self.assertEqual(indexed, [source.resolve()])
+
+
+if __name__ == "__main__":
+    unittest.main()