From 9d37c81f8b67d94c041617b1e87a49e930963912 Mon Sep 17 00:00:00 2001
From: Brendan Ryan <1572504+brendanjryan@users.noreply.github.com>
Date: Thu, 25 Jun 2026 11:09:43 -0700
Subject: [PATCH] fix: harden generated pages and dev server paths

---
 scripts/gen_index.py    |  7 +++--
 scripts/gen_problems.py |  7 +++--
 scripts/serve.py        | 63 ++++++++++++++++++++++++++++++-----------
 scripts/test_serve.py   | 26 +++++++++++++++++
 4 files changed, 82 insertions(+), 21 deletions(-)
 create mode 100644 scripts/test_serve.py

diff --git a/scripts/gen_index.py b/scripts/gen_index.py
index 877fd3f2..ed575c0f 100644
--- a/scripts/gen_index.py
+++ b/scripts/gen_index.py
@@ -11,7 +11,7 @@
 import os
 import re
 
-from jinja2 import Environment, FileSystemLoader
+from jinja2 import Environment, FileSystemLoader, select_autoescape
 
 ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
 SPECS_DIR = os.path.join(ROOT, "specs")
@@ -139,7 +139,10 @@ def build_tree(specs):
 
 
 def main():
-    env = Environment(loader=FileSystemLoader(TEMPLATES_DIR), autoescape=False)
+    env = Environment(
+        loader=FileSystemLoader(TEMPLATES_DIR),
+        autoescape=select_autoescape(default_for_string=True, default=True),
+    )
     tpl = env.get_template("index.html")
 
     specs = collect_specs()
diff --git a/scripts/gen_problems.py b/scripts/gen_problems.py
index ee9cdaa8..c9db6585 100644
--- a/scripts/gen_problems.py
+++ b/scripts/gen_problems.py
@@ -12,7 +12,7 @@
 import json
 import os
 
-from jinja2 import Environment, FileSystemLoader
+from jinja2 import Environment, FileSystemLoader, select_autoescape
 
 ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
 TEMPLATES_DIR = os.path.join(ROOT, "pages", "templates")
@@ -86,7 +86,10 @@ def make_example(slug, title, http_status):
 
 
 def main():
-    env = Environment(loader=FileSystemLoader(TEMPLATES_DIR), autoescape=False)
+    env = Environment(
+        loader=FileSystemLoader(TEMPLATES_DIR),
+        autoescape=select_autoescape(default_for_string=True, default=True),
+    )
     problem_tpl = env.get_template("problem.html")
     index_tpl = env.get_template("problems_index.html")
 
diff --git a/scripts/serve.py b/scripts/serve.py
index c94744bd..7e0110c0 100755
--- a/scripts/serve.py
+++ b/scripts/serve.py
@@ -2,34 +2,63 @@
 """Dev server that serves pages/ and artifacts/ directly. No copying needed."""
 
 import http.server
-import os
+import posixpath
 import sys
+import urllib.parse
+from pathlib import Path
 
-ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
-PAGES = os.path.join(ROOT, "pages")
-ARTIFACTS = os.path.join(ROOT, "artifacts")
+ROOT = Path(__file__).resolve().parents[1]
+PAGES = ROOT / "pages"
+ARTIFACTS = ROOT / "artifacts"
+INVALID_PATH = PAGES / "__invalid_path__"
+
+
+def clean_request_parts(path):
+    path = urllib.parse.unquote(path.split("?", 1)[0].split("#", 1)[0])
+    if ".." in path.lstrip("/").split("/"):
+        return None
+
+    path = posixpath.normpath(path).lstrip("/")
+    if path in ("", "."):
+        return []
+
+    parts = path.split("/")
+    if any(part in ("", ".", "..") for part in parts):
+        return None
+    return parts
+
+
+def resolve_under(root, *parts):
+    candidate = root.joinpath(*parts).resolve()
+    try:
+        candidate.relative_to(root)
+    except ValueError:
+        return INVALID_PATH
+    return candidate
 
 
 class Handler(http.server.SimpleHTTPRequestHandler):
     def translate_path(self, path):
-        path = path.split("?", 1)[0].split("#", 1)[0]
-        path = path.strip("/")
-        if path == "" or path == "index.html":
-            return os.path.join(PAGES, "index.html")
-        if path.startswith("fonts/"):
-            return os.path.join(PAGES, path)
-        if path.startswith("problems"):
+        parts = clean_request_parts(path)
+        if parts is None:
+            return str(INVALID_PATH)
+
+        if parts == [] or parts == ["index.html"]:
+            return str(PAGES / "index.html")
+        if parts[0] == "fonts":
+            return str(resolve_under(PAGES, *parts))
+        if parts[0] == "problems":
             # Serve problem pages with directory-style index.html
-            candidate = os.path.join(PAGES, path)
-            if os.path.isdir(candidate):
-                return os.path.join(candidate, "index.html")
-            return candidate
-        return os.path.join(ARTIFACTS, path)
+            candidate = resolve_under(PAGES, *parts)
+            if candidate.is_dir():
+                return str(candidate / "index.html")
+            return str(candidate)
+        return str(resolve_under(ARTIFACTS, *parts))
 
 
 if __name__ == "__main__":
     port = int(sys.argv[1]) if len(sys.argv) > 1 else 8765
-    server = http.server.HTTPServer(("", port), Handler)
+    server = http.server.HTTPServer(("127.0.0.1", port), Handler)
     print(f"Serving at http://localhost:{port}")
     print(f"  pages/    -> index.html")
     print(f"  artifacts/ -> everything else")
diff --git a/scripts/test_serve.py b/scripts/test_serve.py
new file mode 100644
index 00000000..ecddb2c7
--- /dev/null
+++ b/scripts/test_serve.py
@@ -0,0 +1,26 @@
+"""Tests for the local site server path handling."""
+
+import pytest
+
+from serve import ARTIFACTS, INVALID_PATH, PAGES, clean_request_parts, resolve_under
+
+
+@pytest.mark.parametrize(
+    ("path", "parts"),
+    [
+        ("/draft-example.html?download=1#section", ["draft-example.html"]),
+        ("/%2e%2e/LICENSE.md", None),
+        ("/problems/../LICENSE.md", None),
+        ("/problems/session/channel-not-found/", ["problems", "session", "channel-not-found"]),
+    ],
+)
+def test_clean_request_parts(path, parts):
+    assert clean_request_parts(path) == parts
+
+
+def test_resolve_under_accepts_child_path():
+    assert resolve_under(PAGES, "problems") == PAGES / "problems"
+
+
+def test_resolve_under_rejects_parent_escape():
+    assert resolve_under(ARTIFACTS, "..", "LICENSE.md") == INVALID_PATH