Improve clone-stats workflow: dynamic repo, branch handling, and robust fetch/parsing

[Sendipad]

Motivation

• Make the clone statistics workflow repository-agnostic by removing a hardcoded repo name and ensure the stats branch exists before updates.
• Improve robustness of the network fetch and JSON parsing so the action does not fail on missing values or silent curl errors.

Description

• Add an early step to create/switch to the stats branch with git checkout -B stats before fetching or writing files.
• Use a dynamic API endpoint with "https://api.github.com/repos/${{ github.repository }}/traffic/clones" instead of a hardcoded repository path.
• Change curl to use -sS to surface errors and update jq parsing to jq -r '.count // 0' to default to 0 when the field is absent.
• Create a stats directory, write stats/clones.json, and update git add/commit to include that path while removing the redundant checkout in the commit step.

Testing

• No automated tests were executed for this workflow file change.

---

Codex Task