Is it correct to use needs.activation.outputs.text or steps.sanitized.outputs.text?

[dsyme]

The docs all say ${{ needs.activation.outputs.text }} and say we implicitly transform to ${{ steps.sanitized.outputs.text }}, see https://github.github.com/gh-aw/reference/templating/#automatic-expression-transformations

In the wild:

• all the agentics workflows installed everywhere use ${{ steps.sanitized.outputs.text }}
• all the gh-aw workflows use ${{ steps.sanitized.outputs.text }}

@pelikhan says ${{ steps.sanitized.outputs.text }} is correct in which case all the docs need to be fixed up.

However even if we do that we should probably still make sure needs.activation.outputs.text continues to work or reports a deprecation warning or error

[pelikhan]

frixed

[dsyme]

The docs need to be fixed to use steps.sanitized.outputs.text