Layer autoresearch

Adapted from karpathy/autoresearch program.md. You are an autonomous researcher making the TDK runtime layers cheaper without removing features.

Setup (with the user, once)

1. Agree a run tag (e.g. sep25). Create git checkout -b autoresearch/<tag> from main. The branch must be new.
2. Warn the user: this workspace has an external tool that auto-commits and pushes working-tree edits. Ask them to pause it for the run, because experiment resets must not reach origin/main.
3. Read the in-scope files in full:
   • Editable: engine/topologies/platform/docker/layers/*.star, engine/topologies/platform/docker/layers/prisma/*.star, engine/topologies/platform/docker/build/golden_image_dockerfile.star, engine/topologies/platform/docker/config/healthcheck.star.
   • Read-only ground truth: scripts/autoresearch/layer-eval.sh, scripts/benchmark/container-scale.ts, tests/tilt-engine/test_runtime_container_footprint.py, and the spec openspec/specs/runtime-container-footprint (or its change under openspec/changes/).
4. Check that ../tdk-erp-system exists with built images (docker images | grep finance_). If not, ask the human to run tdk up there once.
5. Create an untracked autoresearch-results.tsv with the header commit mem_mib image_mib status description.
6. Confirm with the user, then start.

Experiment

Run scripts/autoresearch/layer-eval.sh > run.log 2>&1 (about 5-10 minutes). Read the results with grep -A20 '^---' run.log.