GO Benchmark / Case 01 / 2026-11-15–16 (Shanghai)
Retrospective launch-team self-assessment · 0.1-retrospective-pilot

Unzip into a folder, open a terminal there and run with Python 3.9+:
  python3 evaluate.py manifest.json my-result
Use a new or empty output directory. Compare my-result/report.json with expected-report.json.

Expected: estimated_level G1; 21 AI, 8 mixed, 1 human, 2 not_done.
Excerpts are redacted extracts from the project conversation. Hashes apply to public files.
The same team prepared and evaluated the case. Offline human work may be absent.
For your own assessment see AGENT.md and METHOD.md in the general kit at https://gobenchmark.ai/kit/kit.zip
