lili 3a117c1ac4 feat(cheap-worker): U2 compare_node 三品类对照 + 金标 spec 注入(生成/play 拆步) ...

cheap_studio.run_studio 加 run_gates 开关:gen-only(=False)跳过内部
ensure_play_spec+九门 play,让对照方在生成与 play 之间注入金标 spec。

compare_node 扩三品类×n:两路各 generation-only → 强制覆写金标 spec 到
staged _wg1-gen/<id>/play-spec.json(play.cdp.cjs 真实读取处,非 evidence)
→ 各自单独 play,把驱动器从对照变量摘掉(对照公平)。judge_genre 两层判据
+ 逐品类绝对地板 + 双低标红;按品类聚合产框架等价证据(金标 spec 下);
前缀隔离串行;报告分批次不覆盖。

修双评审:金标注入 staged 真路径(Codex C1/C4)、判据绝对地板防双低假等价
(Opus O4/F2)。test_compare_multi 13/13;回归 test_compare 8/8 零破。

2026-06-26 10:49:47 -07:00

..

fixtures/golden-specs

feat(cheap-worker): U1 三品类金标 play-spec fixtures

2026-06-26 10:43:39 -07:00

tests

feat(cheap-worker): U2 compare_node 三品类对照 + 金标 spec 注入(生成/play 拆步)

2026-06-26 10:49:47 -07:00

_bootstrap.py

feat(cheap-worker): 便宜档 Node→Python 重写进 AgentScope spike — 范式可移植坐实

2026-06-26 02:42:14 -07:00

cheap_roles.py

feat(cheap-worker): 便宜档 Node→Python 重写进 AgentScope spike — 范式可移植坐实

2026-06-26 02:42:14 -07:00

cheap_run.py

feat(cheap-worker): 便宜档 Node→Python 重写进 AgentScope spike — 范式可移植坐实

2026-06-26 02:42:14 -07:00

cheap_studio.py

feat(cheap-worker): U2 compare_node 三品类对照 + 金标 spec 注入(生成/play 拆步)

2026-06-26 10:49:47 -07:00

cheap_toolkit.py

feat(cheap-worker): 便宜档 Node→Python 重写进 AgentScope spike — 范式可移植坐实

2026-06-26 02:42:14 -07:00

compare_node.py

feat(cheap-worker): U2 compare_node 三品类对照 + 金标 spec 注入(生成/play 拆步)

2026-06-26 10:49:47 -07:00

requirements.txt

feat(cheap-worker): 便宜档 Node→Python 重写进 AgentScope spike — 范式可移植坐实

2026-06-26 02:42:14 -07:00