Fix: Handle null description in GitHub label schema

[MatrixDai]

Description

#693
When a GitHub issue has labels with null descriptions (rather than missing descriptions), the GitHub MCP server fails with a validation error: "MCP error -32603: Invalid arguments: labels.0 description: Expected string, received null"

Server Details

• Server: github
• Changes to: tools get_issue

Motivation and Context

Fixes #693

How Has This Been Tested?

Test via Cursor, now the MCP tool call get_issue works with label description null

Breaking Changes

NA

Types of changes

• Bug fix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to change)
• Documentation update

Checklist

• I have read the MCP Protocol Documentation
• My changes follows MCP security best practices
• I have updated the server's README accordingly
• I have tested this with an LLM client
• My code follows the repository's style guidelines
• New and existing tests pass locally
• I have added appropriate error handling
• I have documented all environment variables and configuration options

Additional context

NA

[alxleo]

@tadasant Is this safe to merge? This bug is blocking the usage of this Github MCP in tools like Codium Windsurf when using it for calling "list_pull_requests" and "get_pull_request" functions where a particular PR returns a description with null value.

[tadasant]

@alxleo @MatrixDai I'd be happy to merge this if we can get CI to pass. I know your change isn't breaking it, just want CI to be in a good place before merging a code change. Likely is just this #872 + equivalent for Python?

[olaservo]

Edit: sorry, realized later this was a python CI failure not node packages, plus saw the comment below.

[tadasant]

@olaservo I took a closer look to see what might've happened -- it seems like this commit was the only code change in the commit history between working and broken CI.

It looks like this was the PR I merged that included that commit. Weirdly, it wasn't on the git diff / change I reviewed. I think that's just some quirk resulting from a web of merge commits. And it's on a TS file anyway, vs. it was Python that started failing. All to say, I don't think any code change caused this current break in CI.

Need to spend more time figuring out what's wrong here before I can propose a reliable solution, just flagging that I don't think it's a matter of missing something on a code change PR. Something more systematic with how CI is set up. In the meantime I think asking folks making code changes to manually make sure CI is passing is a temporary workaround.

[tadasant]

FYI fixing CI here #1104 then will get this in after