Merge pull request #11 from currentslab/agent/0.1.3-followup

fix: 0.1.3 follow-ups from post-0.1.2 review

Author: Autumn-Atlas

CHANGELOG.md

@@ -1,5 +1,22 @@
 # Changelog
 
+## 0.1.3
+
+### Fixed
+- Restored documented `YYYY-MM-DD` string support in `search()` (broken in
+  0.1.2 when naive datetimes started being rejected); date-only strings are
+  formatted as UTC midnight. Naive timestamp strings (e.g.
+  `2024-01-15T10:00:00`) remain rejected as ambiguous.
+- `examples/_example_utils.safe_markdown_url` now validates with `urlsplit`
+  (scheme + hostname required) and percent-encodes angle brackets,
+  parentheses, and spaces, closing a Markdown/HTML injection path where a
+  `https://...><script>...` URL passed validation.
+- `CurrentsAPI(domain=...)` now requires explicit `allow_custom_domain=True`
+  for any domain other than the default, since a custom domain receives your
+  API key on every request.
+- `examples/source_linked_briefing` catches `CurrentsAPIError` and prints a
+  clean error message (with HTTP status) instead of a traceback.
+
 ## 0.1.2
 
 ### Fixed

currentsapi/__init__.py

@@ -1,7 +1,7 @@
 import sys
 
 __project__ = "currentsapi"
-__version__ = "0.1.2"
+__version__ = "0.1.3"
 __repo__ = "https://github.com/currentslab/currentsapi-python"
 
 from .client import CurrentsAPI

currentsapi/client.py

@@ -45,9 +45,16 @@ def __init__(
         domain=constants.DOMAIN,
         version=constants.VERSION,
         timeout=30,
+        allow_custom_domain=False,
     ):
         if not isinstance(api_key, str):
             raise ValueError("api_key must be a string")
+        if domain != constants.DOMAIN and not allow_custom_domain:
+            raise ValueError(
+                "Passing a custom domain forwards your API key to that host. "
+                "If this is intentional (e.g. testing), pass "
+                "allow_custom_domain=True."
+            )
         self.api_key = ApiAuth(api_key)
         self.latest_endpoint = constants.LATEST_NEWS_URL % (domain, version)
         self.search_endpoint = constants.SEARCH_URL % (domain, version)
@@ -153,6 +160,10 @@ def available_category(self):
     @staticmethod
     def _parse_date(date_value, param_name):
         if isinstance(date_value, str):
+            try:
+                return datetime.date.fromisoformat(date_value)
+            except ValueError:
+                pass
             try:
                 return parser.parse(date_value)
             except (OverflowError, ValueError) as exc:

examples/_example_utils.py

@@ -1,6 +1,7 @@
 """Shared helpers for the shipped Currents API examples."""
 
 import re
+from urllib.parse import urlsplit
 
 
 _MARKDOWN_ESCAPES = re.compile(r"([\\`*_\[\]()<>#!|{}])")
@@ -14,19 +15,25 @@ def escape_markdown_text(value):
 def safe_markdown_url(url, allowed_schemes=("http", "https")):
     """Return a URL safe to embed in Markdown, or ``None`` if unsafe.
 
-    Rejects empty values, non-string values, unknown/unsafe schemes
-    (e.g. ``javascript:``), and whitespace/control characters that could
-    break out of a Markdown link target.
+    Requires an explicit allowed scheme (default http/https) and a hostname;
+    percent-encodes characters that could break out of ``[..](..)`` or
+    ``<..>`` Markdown targets (parentheses, angle brackets, spaces,
+    control characters).
     """
     if not isinstance(url, str):
         return None
     candidate = url.strip()
-    if not candidate:
+    if not candidate or any(ord(ch) < 0x20 for ch in candidate):
         return None
-    if any(ch.isspace() or ord(ch) < 0x20 for ch in candidate):
+    parts = urlsplit(candidate)
+    if parts.scheme.lower() not in allowed_schemes or not parts.netloc:
         return None
-    match = re.match(r"^([A-Za-z][A-Za-z0-9+.-]*):", candidate)
-    if not match or match.group(1).lower() not in allowed_schemes:
+    if any(ch.isspace() for ch in candidate):
         return None
-    # Escape parentheses so the URL cannot break out of [..](..) syntax.
-    return candidate.replace("(", "%28").replace(")", "%29")
+    return (
+        candidate.replace("<", "%3C")
+        .replace(">", "%3E")
+        .replace("(", "%28")
+        .replace(")", "%29")
+        .replace(" ", "%20")
+    )

examples/source_linked_briefing/briefing.py

@@ -13,6 +13,14 @@
 
 sys.path.pop(0)
 
+try:
+    from currentsapi.client import CurrentsAPIError
+except ImportError:  # pragma: no cover - currentsapi not installed
+    class CurrentsAPIError(Exception):
+        """Placeholder when currentsapi is unavailable; never raised."""
+
+        pass
+
 
 def parse_args():
     parser = argparse.ArgumentParser(description=__doc__)
@@ -135,6 +143,11 @@ def main():
         validate_response(response)
         generated_at = resolve_generated_at(args, response)
         markdown, structured = build_output(response, generated_at)
+    except CurrentsAPIError as exc:
+        status = exc.status if exc.status is not None else "unknown"
+        raise SystemExit(
+            "error: Currents API request failed (HTTP {}): {}".format(status, exc)
+        )
     except (OSError, ValueError, json.JSONDecodeError) as exc:
         raise SystemExit("error: {}".format(exc))
 

tests/test_briefing_security.py

@@ -0,0 +1,88 @@
+"""Security regression tests for the source-linked briefing example."""
+
+import importlib.util
+import sys
+from pathlib import Path
+
+EXAMPLES_DIR = Path(__file__).resolve().parent.parent / "examples"
+sys.path.insert(0, str(EXAMPLES_DIR))
+
+from _example_utils import escape_markdown_text, safe_markdown_url  # noqa: E402
+
+_spec = importlib.util.spec_from_file_location(
+    "briefing", EXAMPLES_DIR / "source_linked_briefing" / "briefing.py"
+)
+briefing = importlib.util.module_from_spec(_spec)
+_spec.loader.exec_module(briefing)
+
+
+class TestSafeMarkdownUrl:
+    def test_javascript_scheme_rejected(self):
+        assert safe_markdown_url("javascript:alert(1)") is None
+
+    def test_data_scheme_rejected(self):
+        assert safe_markdown_url("data:text/html,<script>alert(1)</script>") is None
+
+    def test_scheme_relative_rejected(self):
+        assert safe_markdown_url("//attacker.example/x") is None
+
+    def test_missing_netloc_rejected(self):
+        assert safe_markdown_url("https://") is None
+
+    def test_angle_brackets_percent_encoded(self):
+        out = safe_markdown_url("https://example.com><script>x</script>")
+        assert "<" not in out and ">" not in out
+        assert out == "https://example.com%3E%3Cscript%3Ex%3C/script%3E"
+
+    def test_parens_percent_encoded(self):
+        assert safe_markdown_url("https://example.com/x(1)") == "https://example.com/x%281%29"
+
+    def test_normal_url_unchanged(self):
+        assert safe_markdown_url("https://example.com/a?b=1&c=2") == "https://example.com/a?b=1&c=2"
+
+    def test_non_string_rejected(self):
+        assert safe_markdown_url(None) is None
+        assert safe_markdown_url(123) is None
+
+
+class TestBriefingOutputSanitization:
+    def test_hostile_title_cannot_form_link(self):
+        art = {
+            "title": "[click here](javascript:alert(1))",
+            "description": "",
+            "url": "",
+            "published": "2026-01-01",
+        }
+        md, _ = briefing.build_output(
+            {"status": "ok", "news": [art]}, generated_at="2026-01-01T00:00:00Z"
+        )
+        assert "](javascript:" not in md
+
+    def test_hostile_url_cannot_break_out(self):
+        art = {
+            "title": "t",
+            "description": "",
+            "url": "https://example.com><script>location.href='//attacker.example'</script>",
+            "published": "2026-01-01",
+        }
+        md, _ = briefing.build_output(
+            {"status": "ok", "news": [art]}, generated_at="2026-01-01T00:00:00Z"
+        )
+        assert "<script>" not in md
+
+    def test_clean_url_rendered(self):
+        art = {
+            "title": "t",
+            "description": "",
+            "url": "https://example.com/story",
+            "published": "2026-01-01",
+        }
+        md, _ = briefing.build_output(
+            {"status": "ok", "news": [art]}, generated_at="2026-01-01T00:00:00Z"
+        )
+        assert "<https://example.com/story>" in md
+
+    def test_escape_markdown_text_neutralizes_link_syntax(self):
+        escaped = escape_markdown_text("[x](javascript:alert(1))")
+        assert "\\[x\\]" in escaped
+        assert "\\(javascript:" in escaped

tests/test_client.py

@@ -2,6 +2,7 @@
 import unittest
 from unittest.mock import Mock, patch
 
+from currentsapi import constants
 from currentsapi import CurrentsAPI
 from currentsapi.client import CurrentsAPIError
 
@@ -34,7 +35,7 @@ def test_urls_setup(self):
             "https://api.currentsapi.services/v1/available/categories",
         )
 
-        api = CurrentsAPI("dummy_key", "localhost", "v0")
+        api = CurrentsAPI("dummy_key", "localhost", "v0", allow_custom_domain=True)
         self.assertEqual(api.latest_endpoint, "https://localhost/v0/latest-news")
         self.assertEqual(api.search_endpoint, "https://localhost/v0/search")
 
@@ -262,5 +263,29 @@ def test_exception_accepts_non_dict_payload(self):
         self.assertIsNone(exc.code)
         self.assertEqual(str(exc), "Unknown API error")
 
+    @patch("currentsapi.client.requests.get")
+    def test_date_only_string_accepted(self, mock_get):
+        mock_get.return_value = Mock(status_code=200, json=Mock(return_value={"status": "ok"}))
+        api = CurrentsAPI("key")
+        api.search(start_date="2024-01-15", end_date="2024-06-30")
+        kwargs = mock_get.call_args.kwargs
+        self.assertEqual(kwargs["params"]["start_date"], "2024-01-15T00:00:00Z")
+        self.assertEqual(kwargs["params"]["end_date"], "2024-06-30T00:00:00Z")
+
+    @patch("currentsapi.client.requests.get")
+    def test_naive_timestamp_string_rejected(self, mock_get):
+        api = CurrentsAPI("key")
+        with self.assertRaises(ValueError):
+            api.search(start_date="2024-01-15T10:00:00")
+
+    def test_custom_domain_requires_opt_in(self):
+        with self.assertRaises(ValueError) as ctx:
+            CurrentsAPI("key", domain="attacker.example")
+        self.assertIn("allow_custom_domain", str(ctx.exception))
+        api = CurrentsAPI("key", domain="attacker.example", allow_custom_domain=True)
+        self.assertEqual(api.latest_endpoint, "https://attacker.example/v1/latest-news")
+        api = CurrentsAPI("key", domain=constants.DOMAIN)
+        self.assertEqual(api.latest_endpoint, "https://api.currentsapi.services/v1/latest-news")
+
 if __name__ == "__main__":
     unittest.main()