Enterprise agent platform for building, deploying, and operating AI teammates.
"""Tests for xagent.web.config module, focusing on security functions."""
import pytest
from xagent.web.config import (
FILE_STORAGE_URL_BASE,
MAX_COLLECTION_NAME_BYTES,
MAX_COLLECTION_NAME_LENGTH,
get_file_url,
sanitize_path_component,
)
class TestSanitizePathComponent:
"""Test sanitize_path_component function for path traversal prevention."""
def test_valid_collection_names(self):
"""Test that valid collection names are accepted."""
valid_names = [
"my_collection",
"my collection",
"collection-123",
"test123",
"a",
"A",
"collection_name_123",
"a" * MAX_COLLECTION_NAME_LENGTH, # Maximum length
]
for name in valid_names:
result = sanitize_path_component(name, "collection")
assert result == name
assert isinstance(result, str)
def test_path_traversal_attacks(self):
"""Test that path traversal attacks are rejected."""
malicious_names = [
"../../../etc/passwd",
"..\\..\\..\\windows\\system32",
"/etc/passwd",
"collection/../other",
"collection\\..\\other",
"../collection",
"collection/",
"collection\\",
"/collection",
"\\collection",
]
for name in malicious_names:
with pytest.raises(
ValueError, match="path separators or invalid characters"
):
sanitize_path_component(name, "collection")
def test_empty_and_whitespace(self):
"""Test that empty and whitespace-only names are rejected."""
invalid_names = [
"",
" ",
"\t",
"\n",
"\r\n",
]
for name in invalid_names:
with pytest.raises(ValueError, match="cannot be empty"):
sanitize_path_component(name, "collection")
def test_length_limits(self):
"""Test that length limits are enforced."""
# Test minimum length (already tested by empty string, but test edge case)
single_char = "a"
result = sanitize_path_component(single_char, "collection")
assert result == single_char
# Test maximum length
max_length_name = "a" * MAX_COLLECTION_NAME_LENGTH
result = sanitize_path_component(max_length_name, "collection")
assert result == max_length_name
# Test exceeding maximum length
too_long_name = "a" * (MAX_COLLECTION_NAME_LENGTH + 1)
with pytest.raises(
ValueError, match=f"exceeds maximum length of {MAX_COLLECTION_NAME_LENGTH}"
):
sanitize_path_component(too_long_name, "collection")
def test_utf8_byte_length_limit(self):
"""Test that UTF-8 byte-length limits are enforced for Unicode names."""
within_limit_name = "知" * (
MAX_COLLECTION_NAME_BYTES // len("知".encode("utf-8"))
)
assert (
sanitize_path_component(within_limit_name, "collection")
== within_limit_name
)
over_limit_name = "知" * (
(MAX_COLLECTION_NAME_BYTES // len("知".encode("utf-8"))) + 1
)
assert len(over_limit_name) < MAX_COLLECTION_NAME_LENGTH
with pytest.raises(
ValueError,
match=f"exceeds maximum byte length of {MAX_COLLECTION_NAME_BYTES}",
):
sanitize_path_component(over_limit_name, "collection")
def test_invalid_characters(self):
"""Test that invalid characters are rejected."""
invalid_names = [
"collection@name", # @ symbol
"collection#name", # # symbol
"collection$name", # $ symbol
"collection%name", # % symbol
"collection&name", # & symbol
"collection*name", # * symbol
"collection+name", # + symbol
"collection=name", # = symbol
"collection?name", # ? symbol
"collection|name", # | symbol
"collection<name", # < symbol
"collection>name", # > symbol
"collection:name", # : symbol (Windows path separator)
"collection;name", # ; symbol
"collection'name", # ' symbol
'collection"name', # " symbol
"collection[name", # [ symbol
"collection]name", # ] symbol
"collection{name", # { symbol
"collection}name", # } symbol
"collection,name", # , symbol
"collection.name", # . symbol (but multiple dots might be OK if not path traversal)
"collection!name", # ! symbol
"collection~name", # ~ symbol
]
for name in invalid_names:
with pytest.raises(ValueError, match="contains invalid characters"):
sanitize_path_component(name, "collection")
def test_valid_special_characters(self):
"""Test that allowed spaces and special characters are accepted."""
valid_names = [
"collection name",
"collection_name",
"collection-name",
"collection name-123",
"collection_name-123",
"collection-123_name",
"_collection",
"-collection",
"collection_",
"collection-",
]
for name in valid_names:
result = sanitize_path_component(name, "collection")
assert result == name
def test_whitespace_stripping(self):
"""Test that leading/trailing whitespace is stripped before validation."""
# Valid name with whitespace should be stripped and accepted
name_with_whitespace = " my collection "
result = sanitize_path_component(name_with_whitespace, "collection")
assert result == "my collection"
# But if after stripping it becomes invalid, it should be rejected
with pytest.raises(ValueError, match="cannot be empty"):
sanitize_path_component(" ", "collection")
def test_component_type_in_error_messages(self):
"""Test that component_type appears in error messages."""
with pytest.raises(ValueError, match="Invalid collection name"):
sanitize_path_component("", "collection")
with pytest.raises(ValueError, match="Invalid folder name"):
sanitize_path_component("", "folder")
with pytest.raises(ValueError, match="Invalid path name"):
sanitize_path_component("", "path")
def test_path_traversal_after_basename_extraction(self):
"""Test that path traversal is caught even after basename extraction."""
# Path like "../../../etc" becomes "etc" after basename extraction
# But we check if safe_name != name, so it should still be rejected
malicious = "../../../etc"
with pytest.raises(ValueError, match="path separators or invalid characters"):
sanitize_path_component(malicious, "collection")
def test_unicode_and_special_unicode(self):
valid_unicode_names = [
"collection中文",
"collectioné",
"collectionñ",
"示例知识库集合",
"collection١٢٣",
"知识库_123-β",
]
for name in valid_unicode_names:
result = sanitize_path_component(name, "collection")
assert result == name
invalid_unicode_names = [
"collection🚀",
"collectiοn", # Greek omicron mixed with Latin
"cоllection", # Cyrillic o mixed with Latin
"collectionα", # Greek alpha mixed with Latin
]
for name in invalid_unicode_names:
with pytest.raises(ValueError, match="contains"):
sanitize_path_component(name, "collection")
def test_rejects_compatibility_homoglyph_forms(self):
with pytest.raises(ValueError, match="contains invalid characters"):
sanitize_path_component("Agent", "collection")
def test_numeric_only_names(self):
"""Test that numeric-only names are accepted."""
numeric_names = ["123", "0", "999999"]
for name in numeric_names:
result = sanitize_path_component(name, "collection")
assert result == name
def test_mixed_case_names(self):
"""Test that mixed case names are accepted."""
mixed_names = ["MyCollection", "myCollection", "MY_COLLECTION", "My-Collection"]
for name in mixed_names:
result = sanitize_path_component(name, "collection")
assert result == name
class TestGetFileUrl:
def test_get_file_url_encodes_unicode_collection_and_filename(self):
collection_name = "示例知识库集合"
filename = "报告.txt"
url = get_file_url(filename, user_id=7, collection=collection_name)
assert url == (
f"{FILE_STORAGE_URL_BASE}/user_7/"
"%E7%A4%BA%E4%BE%8B%E7%9F%A5%E8%AF%86%E5%BA%93%E9%9B%86%E5%90%88/"
"%E6%8A%A5%E5%91%8A.txt"
)