[Bug]: UTF-8 Chinese text misclassified as binary when sampling splits a multibyte character

Is there an existing issue for the same bug?

• Searched SDK issues for binaryornot and binary UTF-8; found no issue covering this sample-truncation problem.

Bug Description

FileEditor.validate_file() uses binaryornot.is_binary() to detect binary files. Its 1,024-byte sample can split a multibyte UTF-8 character, causing an otherwise valid Chinese Markdown file to be classified as binary and preventing it from being viewed or edited.

Expected Behavior

Valid UTF-8 text without binary control characters should be viewable and editable through the file editor.

Actual Behavior

Raises FileValidationError: File appears to be binary and this file type cannot be read or edited by this tool.

Steps to Reproduce / Minimal Code Sample

Run from the SDK source directory after installing dependencies:

uv sync --frozen --dev
uv run python - <<'PY'
from pathlib import Path
from tempfile import TemporaryDirectory
from openhands.tools.file_editor.editor import FileEditor

with TemporaryDirectory() as directory:
    path = Path(directory) / "report.md"
    # Chinese text encoded as escapes to keep this description in English.
    path.write_text("\u66f4\u6df1\u4e00\u5c42\u7684\u5f52\u5c5e\u611f\u3002" * 200, encoding="utf-8")
    result = FileEditor()(command="view", path=str(path))
    assert not result.is_error
PY

For this content, the first 1,023 or 1,026 bytes decode successfully as UTF-8, and is_binary_string() returns False. At 1,024 or 1,025 bytes, the sample ends in a truncated character and the function returns True. The complete file decodes successfully using strict UTF-8 decoding.

Acceptance Criteria

• Valid Chinese text can be viewed and replaced at different sample alignments, with correct full UTF-8 contents after replacement.
• A valid UTF-8 prefix must not allow a file containing subsequent NUL bytes, binary control characters, or invalid UTF-8 bytes to bypass validation.
• Preserve existing file-size limits, image handling, and other file-editor behavior.

Installation Method / SDK Version

Installed from source using uv sync --frozen --dev; reproduced on upstream main b5c8ab950 (SDK 1.47.0).

Python Version / Operating System / Model

Python 3.13.7, Linux; no model calls required.

Additional Context

Related to, but distinct from, PR #4784: that PR changes encoding selection in EncodingManager. This issue occurs in the binary rejection check in validate_file() and requires a separate fix.

Start at FileEditor.validate_file() in openhands.tools.file_editor.editor and run the provided UTF-8 reproduction after installing dependencies with uv sync --frozen --dev. Add focused coverage for different sample alignments, then verify that valid UTF-8 remains editable while later NUL bytes, binary controls, or invalid UTF-8 still fail validation and existing size and image behavior is preserved.