feat: update grading workflows and refresh tool versions

[FidelusAleksander]

Summary

Changed the way we do grading workflows, now using https://github.com/skills/action-keyphrase-checker
Removed skills/response-templates references
Updated all action-text-variables refs to v2 and yaml style vars
Added a tip to approve workflow run triggered by copilot review

Changes

Closes:

Task list

• For workflow changes, I have verified the Actions workflows function as expected.
• For content changes, I have reviewed the style guide.

[Copilot]

Pull Request Overview

This PR updates the grading workflows by replacing older toolkit versions and response template references with the new exercise-toolkit v0.4.0 and updates all action-text-variables references to v2. Additionally, the PR refines workflow steps (e.g. PR description and Copilot review checks) and adjusts comment creation to align with the new tool version.

Reviewed Changes

Copilot reviewed 7 out of 7 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
.github/workflows/4b-copilot-on-github.yml	Updated PR trigger configuration, toolkit version bump, and refined check steps.
.github/workflows/4-copilot-on-github.yml	Updated checkout paths, toolkit version bump, and finish-exercise workflow update.
.github/workflows/3b-copilot-agent-mode.yml	Upgraded exercise-toolkit version from v0.1.0 to v0.4.0.
.github/workflows/3-copilot-edits.yml	Replaced custom checks with action-keyphrase-checker and updated text variables.
.github/workflows/2-first-introduction.yml	Updated response templates to exercise-toolkit templates with new version bump.
.github/workflows/1-preparing.yml	Replaced response template references with exercise-toolkit references (v0.4.0).
.github/workflows/0-start-exercise.yml	Updated start-exercise workflow including toolkit version bump and text variable update.

[chriswblake]

Everything looks good. Just confirm if the trigger for workflow 4b should stay disabled or it is safe to enable now.

[arilivigni]

LGTM

Are we sticking with --edit-last or adding a new comment.

          gh issue comment "$ISSUE_URL" \
            --body-file exercise-toolkit/markdown-templates/step-feedback/checking-work.md \
            --edit-last

[FidelusAleksander]

LGTM

Are we sticking with --edit-last or adding a new comment.

          gh issue comment "$ISSUE_URL" \
            --body-file exercise-toolkit/markdown-templates/step-feedback/checking-work.md \
            --edit-last

I think a combination of both is probably good. Some comments make sense to be updated, some not e.g

• we don't want the result table commented separately every time the grading job fails for a given step, so we use --edit-last

      - name: Create comment - step results
        run: |
          gh issue comment "$ISSUE_URL" \
            --body "$COMMENT_BODY" \
            --edit-last
        env:
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
          COMMENT_BODY: ${{ steps.build-message-step-results.outputs.updated-text }}

• It's probably best that if the grading workflow succeeds we don't override the results table with --edit-last step-finished comment so the exercise taker can see these results, in that case we would create a new comment

        - name: Update comment - step finished
          run: |
            gh issue comment "$ISSUE_URL" \
              --body "$ISSUE_BODY"
          env:
            GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
            ISSUE_BODY: ${{ steps.build-message-step-finish.outputs.updated-text }}

All open to change, and probably standardize/document