Feedback: Assay Action v2 evidence review in GitHub Actions

[Rul1an]

Assay Action v2 is meant to be the small GitHub-native entry point for Assay evidence artifacts.

The shape is deliberately copyable: paste a workflow, verify evidence bundles, get a job summary, SARIF/code-scanning output, and a named assay-reports-${{ github.run_id }} artifact for review.

The useful reviewer question is simple:

What evidence did this run produce, and did this PR introduce findings above the configured threshold?

This is not a dashboard, hosted service, model-correctness claim, or upstream integration claim. It is a thin PR review surface over portable evidence artifacts.

Start here:

• README and workflow: https://github.com/Rul1an/assay-action#readme
• v2 release note: https://github.com/Rul1an/assay-action/releases/tag/v2
• Evidence Receipts in Action: https://github.com/Rul1an/assay/blob/main/docs/notes/EVIDENCE-RECEIPTS-IN-ACTION.md

Useful feedback for this thread:

• Was the workflow copyable without reading the full Assay docs first?
• Was the job summary enough to understand what failed a PR?
• Was the reports artifact name easy to find?
• Did baseline-delta output clarify what changed, or add noise?

For the experimental capability diff preview specifically, use #11 so preview feedback stays separate from the stable v2 action surface.