Fix markdown-fix workflow to commit partial fixes when errors remain

[Copilot]

The @phlexbot format command in PR #290 failed because markdownlint-cli2-action exits with code 1 when errors are found, terminating the workflow before the commit step runs.

Changes

• Add continue-on-error: true to markdownlint step in markdown-fix.yaml
• Add MARKDOWN_FIX_ANALYSIS.md documenting the issue and specific errors in PR Workflow standardization and improvements #290

Details

- name: Run markdownlint
  uses: DavidAnson/markdownlint-cli2-action@07035fd053f7be764496c0f8d8f9f41f98305101
  with:
    globs: |
      ${{ env.local_checkout_path }}/**/*.md
      !${{ env.local_checkout_path }}/**/CHANGELOG.md
    fix: true
  continue-on-error: true  # Allows commit step to run even when errors remain

This matches the existing pattern in markdown-check.yaml line 84. The workflow now commits whatever fixes markdownlint can make automatically, even when some errors require manual intervention.

The analysis document provides the exact fixes needed for PR #290's remaining errors (MD040: missing code fence language, MD060: table formatting).

Original prompt

A @phlexbot format command to fix markdown errors in PR 290 was unsuccessful, per https://github.com/Framework-R-D/phlex/actions/runs/21736866948/job/62703615091. Please investigate and fix if possible.

---

💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.