Merge branch 'main' of github.com:NousResearch/hermes-agent into feat/ink-refactor

tests/hermes_cli/test_non_ascii_credential.py

@@ -0,0 +1,83 @@
+"""Tests for non-ASCII credential detection and sanitization.
+
+Covers the fix for issue #6843 — API keys containing Unicode lookalike
+characters (e.g. ʋ U+028B instead of v) cause UnicodeEncodeError when
+httpx tries to encode the Authorization header as ASCII.
+"""
+
+import os
+import sys
+import tempfile
+
+import pytest
+
+from hermes_cli.config import _check_non_ascii_credential
+
+
+class TestCheckNonAsciiCredential:
+    """Tests for _check_non_ascii_credential()."""
+
+    def test_ascii_key_unchanged(self):
+        key = "sk-proj-" + "a" * 100
+        result = _check_non_ascii_credential("TEST_API_KEY", key)
+        assert result == key
+
+    def test_strips_unicode_v_lookalike(self, capsys):
+        """The exact scenario from issue #6843: ʋ instead of v."""
+        key = "sk-proj-abc" + "ʋ" + "def"  # \u028b
+        result = _check_non_ascii_credential("OPENROUTER_API_KEY", key)
+        assert result == "sk-proj-abcdef"
+        assert "ʋ" not in result
+        # Should print a warning
+        captured = capsys.readouterr()
+        assert "non-ASCII" in captured.err
+
+    def test_strips_multiple_non_ascii(self, capsys):
+        key = "sk-proj-aʋbécd"
+        result = _check_non_ascii_credential("OPENAI_API_KEY", key)
+        assert result == "sk-proj-abcd"
+        captured = capsys.readouterr()
+        assert "U+028B" in captured.err  # reports the char
+
+    def test_empty_key(self):
+        result = _check_non_ascii_credential("TEST_KEY", "")
+        assert result == ""
+
+    def test_all_ascii_no_warning(self, capsys):
+        result = _check_non_ascii_credential("KEY", "all-ascii-value-123")
+        assert result == "all-ascii-value-123"
+        captured = capsys.readouterr()
+        assert captured.err == ""
+
+
+class TestEnvLoaderSanitization:
+    """Tests for _sanitize_loaded_credentials in env_loader."""
+
+    def test_strips_non_ascii_from_api_key(self, monkeypatch):
+        from hermes_cli.env_loader import _sanitize_loaded_credentials
+
+        monkeypatch.setenv("OPENROUTER_API_KEY", "sk-proj-abcʋdef")
+        _sanitize_loaded_credentials()
+        assert os.environ["OPENROUTER_API_KEY"] == "sk-proj-abcdef"
+
+    def test_strips_non_ascii_from_token(self, monkeypatch):
+        from hermes_cli.env_loader import _sanitize_loaded_credentials
+
+        monkeypatch.setenv("DISCORD_BOT_TOKEN", "tokénvalue")
+        _sanitize_loaded_credentials()
+        assert os.environ["DISCORD_BOT_TOKEN"] == "toknvalue"
+
+    def test_ignores_non_credential_vars(self, monkeypatch):
+        from hermes_cli.env_loader import _sanitize_loaded_credentials
+
+        monkeypatch.setenv("MY_UNICODE_VAR", "héllo wörld")
+        _sanitize_loaded_credentials()
+        # Not a credential suffix — should be left alone
+        assert os.environ["MY_UNICODE_VAR"] == "héllo wörld"
+
+    def test_ascii_credentials_untouched(self, monkeypatch):
+        from hermes_cli.env_loader import _sanitize_loaded_credentials
+
+        monkeypatch.setenv("OPENAI_API_KEY", "sk-proj-allascii123")
+        _sanitize_loaded_credentials()
+        assert os.environ["OPENAI_API_KEY"] == "sk-proj-allascii123"

tests/run_agent/test_unicode_ascii_codec.py

@@ -142,6 +142,33 @@ class TestSurrogateVsAsciiSanitization:
         assert _sanitize_messages_surrogates(messages) is False
 
 
+class TestApiKeyNonAsciiSanitization:
+    """Tests for API key sanitization in the UnicodeEncodeError recovery.
+
+    Covers the root cause of issue #6843: a non-ASCII character (ʋ U+028B)
+    in the API key causes httpx to fail when encoding the Authorization
+    header as ASCII.  The recovery block must strip non-ASCII from the key.
+    """
+
+    def test_strip_non_ascii_from_api_key(self):
+        """_strip_non_ascii removes ʋ from an API key string."""
+        key = "sk-proj-abc" + "ʋ" + "def"
+        assert _strip_non_ascii(key) == "sk-proj-abcdef"
+
+    def test_api_key_at_position_153(self):
+        """Reproduce the exact error: ʋ at position 153 in 'Bearer <key>'."""
+        key = "sk-proj-" + "a" * 138 + "ʋ" + "bcd"
+        auth_value = f"Bearer {key}"
+        # This is what httpx does — and it fails:
+        with pytest.raises(UnicodeEncodeError) as exc_info:
+            auth_value.encode("ascii")
+        assert exc_info.value.start == 153
+        # After sanitization, it should work:
+        sanitized_key = _strip_non_ascii(key)
+        sanitized_auth = f"Bearer {sanitized_key}"
+        sanitized_auth.encode("ascii")  # should not raise
+
+
 class TestSanitizeToolsNonAscii:
     """Tests for _sanitize_tools_non_ascii."""