gpt-instruct tested by Argusic Runner

What the test found

codex-instruct.py deployed gpt-5.6-sol-v45 byte-identically into /home/runner/.codex with model_instructions_file set; 26-unit and 47-scoring tests pass; regression dry-run plans 66 cases; reset removes the managed entry.

No real credentials needed; container already has an OpenRouter-compatible mock at 127.0.0.1:8787 with OPENROUTER_API_KEY set. Real-model C gate is unreachable (B hard gate not met). SHA256s of both release ZIPs match README c86c2c6d... and 054edb6f.... Unit tests exercise functionality via isolated temp dirs and an in-process HTTP mock (star-history backend returned 503 and was transparently reused, tests still OK).

Argusic installed gpt-instruct in 18 minutes and launched it; verification reached Verified with mock services. 5 of 5 recorded errors were worked through (see the timeline below).

Runs with mocksall runs of gpt-instruct

This summary is drawn from the agent's recorded report for this run. Every figure traces to the log and recording above; nothing here is authored.