Handle empty file paths in cucumber pull

[PierrickTassery]

Summary

This keeps the pull workflow from crashing when the API returns an empty file path in data.files.

When the file path is empty, the CLI now generates a fallback filename from the Gherkin Feature: title and writes the file there instead of trying to write the target directory itself.

Refs #70.

Why

check-cucumber pull currently assumes every key in data.files is a writable filename. If the API returns an empty key:

files: {
  "": "Feature: Missing Source Path\n..."
}

then path.join(targetDir, "") resolves to the target directory and fs.writeFileSync fails with:

EISDIR: illegal operation on a directory, open '.'

Changes

• Add a minimal fallback filename generator for empty file paths.
• Use the feature title as the fallback filename base.
• Add a regression test for an empty file path response.

This PR intentionally does not try to solve duplicate source-path collisions; that requires an API-side change because missing feature contents are not present in data.files.

Testing

npm test

Result:

97 passing

[DavertMik]

@DenysKuchma please check if we can come up wiht better implementaton
regexes look fragile, we need to think about it

[DavertMik]

Thanks I will ask our developer to review and merge or provide alternative solution