From 89b7955b8181c1d4172d3102b514d6860cdc509f Mon Sep 17 00:00:00 2001
From: Claude <noreply@anthropic.com>
Date: Fri, 25 Sep 2026 23:07:35 +0000
Subject: [PATCH] Strip bare control characters from CSP report terminal output

_sanitize_terminal() only removed ESC-prefixed ANSI sequences, so bare
C0 controls (CR, LF, BS, BEL, ...), DEL, C1 controls (e.g. 0x9B 8-bit
CSI) and a stray ESC in unauthenticated /__csp__ report fields reached
the operator's terminal, allowing log lines to be overwritten or forged.

Strip all remaining control characters and Unicode bidi overrides after
the ANSI escape pass.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_019b8mWRK1bS7PUszC6SQKGB
---
 mesop/server/static_file_serving.py      | 14 +++++-
 mesop/server/static_file_serving_test.py | 64 ++++++++++++++++++++++++
 2 files changed, 76 insertions(+), 2 deletions(-)

diff --git a/mesop/server/static_file_serving.py b/mesop/server/static_file_serving.py
index 7c520e746..2980cd07e 100644
--- a/mesop/server/static_file_serving.py
+++ b/mesop/server/static_file_serving.py
@@ -59,13 +59,23 @@ def noop():
   re.VERBOSE,
 )
 
+# Matches control characters that are not part of an ESC-prefixed sequence but
+# can still manipulate terminal output (e.g. CR/BS overwrite text, LF forges new
+# log lines, 0x9B is an 8-bit CSI), plus Unicode bidi overrides which can
+# visually reorder text.
+_CONTROL_CHARS_RE = re.compile(
+  r"[\x00-\x1f\x7f-\x9f\u200e\u200f\u202a-\u202e\u2066-\u2069]"
+)
+
 
 def _sanitize_terminal(value: object) -> str:
-  # Prevent ANSI/VT100 terminal escape injection when logging
+  # Prevent terminal escape and control character injection when logging
   # attacker-controlled data (e.g. CSP report fields).
   if not isinstance(value, str):
     value = str(value)
-  return _ANSI_ESCAPE_RE.sub("", value)
+  value = _ANSI_ESCAPE_RE.sub("", value)
+  # Strip any remaining control characters (including a stray ESC).
+  return _CONTROL_CHARS_RE.sub("", value)
 
 
 def configure_static_file_serving(
diff --git a/mesop/server/static_file_serving_test.py b/mesop/server/static_file_serving_test.py
index 8d941268e..3b75bb5b0 100644
--- a/mesop/server/static_file_serving_test.py
+++ b/mesop/server/static_file_serving_test.py
@@ -97,6 +97,41 @@ def test_sanitize_terminal_keeps_plain_text():
   assert _sanitize_terminal(payload) == payload
 
 
+def test_sanitize_terminal_removes_bare_control_characters():
+  # C0 controls (incl. CR/LF/BS/BEL), DEL, C1 controls (e.g. 8-bit CSI) and
+  # a stray ESC are not ESC-prefixed sequences but still affect the terminal.
+  for ch in [
+    "\x00",
+    "\x07",
+    "\x08",
+    "\t",
+    "\n",
+    "\x0b",
+    "\x0c",
+    "\r",
+    "\x0e",
+    "\x0f",
+    "\x1b",
+    "\x7f",
+    "\x85",
+    "\x9b",
+  ]:
+    assert _sanitize_terminal(f"before{ch}after") == "beforeafter", repr(ch)
+
+
+def test_sanitize_terminal_removes_bidi_overrides():
+  rlo, lri, pdi = chr(0x202E), chr(0x2066), chr(0x2069)
+  payload = f"abc{rlo}def{lri}g{pdi}"
+
+  assert _sanitize_terminal(payload) == "abcdefg"
+
+
+def test_sanitize_terminal_keeps_non_ascii_text():
+  payload = "https://例え.jp/café?q=ü"
+
+  assert _sanitize_terminal(payload) == payload
+
+
 def test_sanitize_terminal_coerces_non_str_input():
   assert _sanitize_terminal(123) == "123"
 
@@ -132,6 +167,35 @@ def test_csp_report_sanitizes_ansi_escapes_in_output(capsys):
   assert "evil.example" in output
 
 
+def test_csp_report_strips_control_characters_in_output(capsys):
+  app = Flask(__name__)
+  configure_static_file_serving(
+    app,
+    static_file_runfiles_base="unused",
+    disable_gzip_cache=True,
+  )
+  client = app.test_client()
+
+  response = client.post(
+    "/__csp__",
+    json={
+      "csp-report": {
+        "document-uri": "https://example.com/page\nFAKE: forged line",
+        "blocked-uri": "https://evil.example/x\rINFO: server healthy\x08\x07",
+        "violated-directive": "connect-src",
+      }
+    },
+  )
+
+  assert response.status_code == 204
+  output = capsys.readouterr().out
+  for ch in ["\r", "\x08", "\x07"]:
+    assert ch not in output
+  assert "\nFAKE" not in output
+  assert "https://evil.example/xINFO: server healthy" in output
+  assert "/pageFAKE: forged line" in output
+
+
 if __name__ == "__main__":
   import pytest