fix: detect and strip non-ASCII characters from API keys (#6843)

API keys containing Unicode lookalike characters (e.g. ʋ U+028B instead
of v) cause UnicodeEncodeError when httpx encodes the Authorization
header as ASCII.  This commonly happens when users copy-paste keys from
PDFs, rich-text editors, or web pages with decorative fonts.

Three layers of defense:

1. **Save-time validation** (hermes_cli/config.py):
   _check_non_ascii_credential() strips non-ASCII from credential values
   when saving to .env, with a clear warning explaining the issue.

2. **Load-time sanitization** (hermes_cli/env_loader.py):
   _sanitize_loaded_credentials() strips non-ASCII from credential env
   vars (those ending in _API_KEY, _TOKEN, _SECRET, _KEY) after dotenv
   loads them, so the rest of the codebase never sees non-ASCII keys.

3. **Runtime recovery** (run_agent.py):
   The UnicodeEncodeError recovery block now also sanitizes self.api_key
   and self._client_kwargs['api_key'], fixing the gap where message/tool
   sanitization succeeded but the API key still caused httpx to fail on
   the Authorization header.

Also: hermes_logging.py RotatingFileHandler now explicitly sets
encoding='utf-8' instead of relying on locale default (defensive
hardening for ASCII-locale systems).
This commit is contained in:
Teknium
2026-04-14 17:17:15 -07:00
committed by Teknium
parent 677f1227c3
commit da528a8207
6 changed files with 206 additions and 0 deletions

View File

@@ -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"