Beyond Artifact-Only Evaluation: A Case for Development-Session Attestation (DSA)

Introduction AI-assisted development has dramatically increased implementation speed. At the same time, I think it has made one thing harder: relying on artifact-only evaluation as a sufficient signal of engineering work (repositories, portfolio sites, demos, finished UI screenshots, etc.). Artifacts still matter. Code review, testing, CI, and running systems will remain essential. The problem is not that artifacts became useless. The problem is that artifacts alone often do not answer questions like: How much of the implementation was actually understood? What tools/processes were executed during development? Is the development process itself explainable after the fact? This becomes especially important in situations where implementation output is not the whole story: incident response and quality explanation hiring assessments constrained test environments (e.g., prohibited tools) training and education contexts This is the gap I am trying to frame with DSA (Development-Session Attes