Polish TwinBench for S-tier public launch #8

	name: Harness CI

	on:
	push:
	branches: [main]
	pull_request:
	branches: [main]

	jobs:
	harness-ci:
	runs-on: ubuntu-latest
	steps:
	- name: Checkout
	uses: actions/checkout@v4

	- name: Setup Python 3.10
	uses: actions/setup-python@v5
	with:
	python-version: "3.10"

	- name: Install dependencies
	run: \|
	python -m pip install --upgrade pip
	pip install -r harness/requirements.txt

	- name: Compile harness sources
	run: python -m py_compile harness/*.py

	- name: Runner CLI smoke
	run: python -m harness.runner --help

	- name: Validate v0.2 result artifact schema
	run: \|
	python - <<'PY'
	import json
	from pathlib import Path

	p = Path("results/nullalis-v0.2.json")
	if not p.exists():
	raise SystemExit("results/nullalis-v0.2.json not found")

	data = json.loads(p.read_text())
	required = [
	"benchmark_version",
	"verified_composite_score",
	"projected_composite_score",
	"measured_coverage",
	"coverage_adjusted_verified_score",
	"dimension_verified_scores",
	"dimension_projected_scores",
	"dimension_measured_coverage",
	]
	missing = [k for k in required if k not in data]
	if missing:
	raise SystemExit(f"missing keys: {missing}")
	if data["benchmark_version"] != "0.2":
	raise SystemExit("benchmark_version must be 0.2")
	print("artifact schema ok")
	PY

Provide feedback