Commit Graph

Select branches

Hide Pull Requests

codex/fast-slow-evaluation

main

Mono Color

• e2e7a7ba37 docs: add complete Chinese documentation main wuyang 2026-07-26 21:32:12 +08:00
• 66109235e2 style: adopt unit-01 inspired docs palette wuyang 2026-07-26 21:19:41 +08:00
• a895e04bc3 docs: link portal to published sources wuyang 2026-07-26 21:09:49 +08:00
• 688ddec1fe docs: publish architecture and experiment portal wuyang 2026-07-26 21:03:23 +08:00
• 4eb915283d fix: hide workspace dependency directories wuyang 2026-07-26 20:28:52 +08:00
• 2fbb5698ac fix: persist workspace Python environments wuyang 2026-07-26 20:20:37 +08:00
• bc30a480ad refine model and thinking selectors wuyang 2026-07-26 19:18:28 +08:00
• d95d06233f refactor: use a single agent model path wuyang 2026-07-26 18:50:51 +08:00
• 2e8e81c790 feat: add DeepSeek extreme reasoning tier wuyang 2026-07-26 17:40:51 +08:00
• ade5ea1210 fix: make invalid source retries actionable wuyang 2026-07-26 17:08:27 +08:00
• f64380a0f4 fix: ground Work reports in measurements wuyang 2026-07-26 17:00:23 +08:00
• b49b93dcf7 fix: require downloadable Work reports wuyang 2026-07-26 16:48:28 +08:00
• 5c9f827357 fix: require executable source for Work reports wuyang 2026-07-26 16:35:58 +08:00
• 87689931b5 fix: preserve tool failures under noisy output wuyang 2026-07-26 16:14:26 +08:00
• af141845ed fix: bound Work model generation wuyang 2026-07-26 16:03:59 +08:00
• be7abbd18f fix: report workspace status success wuyang 2026-07-26 15:10:56 +08:00
• e5207a3dda fix: stream Work model tool responses wuyang 2026-07-26 15:10:56 +08:00
• 2c17a74203 fix: retry transient model failures wuyang 2026-07-26 14:46:43 +08:00
• 1216e96038 fix: decode repeated source layout escapes wuyang 2026-07-26 14:37:56 +08:00
• 2dd308cf3b fix: reject invalid and duplicate Work writes wuyang 2026-07-26 14:31:48 +08:00
• f2282c7e7d feat: focus Work tools by task intent wuyang 2026-07-26 14:22:46 +08:00
• ae5f940a5d fix: deduplicate Work execution batches wuyang 2026-07-26 14:11:57 +08:00
• 44fe0b3dd7 fix: normalize escaped Work artifacts wuyang 2026-07-26 14:03:01 +08:00
• f149379cf7 fix: prevent unchanged execution retry loops wuyang 2026-07-26 13:49:30 +08:00
• 8a31a0edb6 fix: recover Work runs after failed verification wuyang 2026-07-26 13:37:51 +08:00
• 2f6f46c2fa fix: brand static application title wuyang 2026-07-26 13:14:30 +08:00
• d30603e0bd fix: preserve SSH config mount in deployment wuyang 2026-07-26 13:07:34 +08:00
• 8cc8b93fc5 fix: package Docker SSH transport wuyang 2026-07-26 13:03:24 +08:00
• 0b4d799216 feat: add remote workspaces and verified deliverables wuyang 2026-07-26 12:58:04 +08:00
• 37f83eaac7 feat: rebuild as multi-user web agent wuyang 2026-07-26 06:52:12 +08:00
• 1e7570e72d fix: bound idle evaluation model streams codex/fast-slow-evaluation wuyang6 2026-07-24 20:08:38 +08:00
• 459bee47a7 fix: recover interrupted evaluation cases wuyang6 2026-07-24 20:04:09 +08:00
• 96809fa5c8 fix: back off evaluation rate limits wuyang6 2026-07-24 17:51:10 +08:00
• 174d1d3d01 feat: add evaluation label mapping wuyang6 2026-07-24 17:34:32 +08:00
• e2bf012f60 feat(skill): improve multi-turn fast-slow routing wuyang6 2026-07-24 17:12:25 +08:00
• a505261bde fix: make skill editor responsive wuyang6 2026-07-24 10:17:05 +08:00
• 6630e82069 feat: add editable skill evaluation versions wuyang6 2026-07-24 10:11:51 +08:00
• ecc02c06b7 refactor: layer fast slow routing knowledge wuyang6 2026-07-23 21:35:18 +08:00
• 6116d10e8f feat: persist single evaluation history wuyang6 2026-07-23 19:46:15 +08:00
• 5be9434ba3 feat: add skill evaluation workbench wuyang6 2026-07-23 19:38:25 +08:00
• 9067914ecc feat: 新增label快慢分流skill全套配置与校验脚本 zhukai8 2026-07-23 18:09:01 +08:00
• 017743a415 update wuyang6 2026-07-23 16:50:47 +08:00
• 0e4f5041f1 docs: add skill evaluation workbench design wuyang6 2026-07-23 16:49:53 +08:00
• 97e1e1dd48 Add versioned annotation submissions wuyang6 2026-07-20 17:28:47 +08:00
• 7cd6e5e464 Add annotation page gallery wuyang6 2026-07-20 10:56:59 +08:00
• b7798e9985 Update screen operation label indexes wuyang6 2026-07-14 15:00:29 +08:00
• ef213c4125 Update screen operation label knowledge wuyang6 2026-07-14 14:57:55 +08:00
• 359b84ba55 Fix auth session race wuyang6 2026-07-13 20:01:21 +08:00
• f5f98688dd Fix compact summary regex replacement wuyang6 2026-07-07 21:21:19 +08:00
• da8b63443b Add shared Jupyter defaults for model training lite wuyang6 2026-07-07 20:00:27 +08:00
• 2a6a7c2780 Enhance model training lite workflow wuyang6 2026-07-07 19:58:00 +08:00
• 8eadc25b19 fix feishu reauth for online docs wuyang6 2026-07-02 14:02:01 +08:00
• 0007d99763 fix session display replay consistency wuyang6 2026-06-30 14:40:20 +08:00
• 84ff22fc65 Handle absolute workspace globs wuyang6 2026-06-24 15:06:12 +08:00
• 3637e2c73a Fix compacted tool history for resumed sessions wuyang6 2026-06-24 14:18:15 +08:00
• 45792c8fd5 Support runtime guidance tool interrupts wuyang6 2026-06-23 15:20:59 +08:00
• 712bccf814 Reject reasoning text in session titles wuyang6 2026-06-16 15:10:41 +08:00
• c8b5b0659d Fix session title refinement after optimistic save wuyang6 2026-06-16 14:56:58 +08:00
• 1d38753d08 Stabilize Jupyter terminal wrapper execution wuyang6 2026-06-15 17:14:54 +08:00
• c795ec1b5d Avoid Jupyter terminal job-control noise wuyang6 2026-06-12 21:38:53 +08:00
• c91f65c856 Filter Jupyter terminal wrapper output wuyang6 2026-06-12 21:32:06 +08:00
• 7422931290 Auto-run queued session inputs wuyang6 2026-06-12 20:17:20 +08:00
• c3caa4a1e3 Keep optimistic user message during replay wuyang6 2026-06-12 20:07:42 +08:00
• 611284b27a Persist theme before hydration wuyang6 2026-06-12 19:58:16 +08:00
• 4a268902bf Persist completed run elapsed time wuyang6 2026-06-12 19:48:45 +08:00
• aac6383fe6 Fix runtime guidance transcript replay wuyang6 2026-06-12 19:41:26 +08:00
• 9237d6782a Widen sidebar session titles wuyang6 2026-06-12 19:27:43 +08:00
• 780f3447b1 Avoid overlapping session title ellipsis wuyang6 2026-06-12 19:23:47 +08:00
• 82b41ce89d Reduce session title action spacing wuyang6 2026-06-12 19:20:47 +08:00
• de6849588e Use ellipsis for session titles wuyang6 2026-06-12 19:17:33 +08:00
• c4f9cca2d7 Tighten session list status slot wuyang6 2026-06-12 18:42:20 +08:00
• cdf27e5bd2 Show runtime guidance in transcript wuyang6 2026-06-12 18:37:04 +08:00
• 5bcc22622e Show session run badges wuyang6 2026-06-12 18:28:00 +08:00
• 4ba8761db2 Add runtime guidance input queue wuyang6 2026-06-12 18:15:52 +08:00
• 00a83ba4c3 Cache session replay on client wuyang6 2026-06-12 17:28:15 +08:00
• ddf06c727d Keep append-only session revisions wuyang6 2026-06-12 16:56:01 +08:00
• 384e92e961 Prevent stale session overwrite on new tasks wuyang6 2026-06-12 16:51:17 +08:00
• a2f39e7312 Fix session cancel and new task routing wuyang6 2026-06-12 16:35:38 +08:00
• 0a3e044c9d Cancel session background processes wuyang6 2026-06-12 16:24:25 +08:00
• 953126e1d3 Refactor session runtime persistence wuyang6 2026-06-12 16:17:58 +08:00
• 77d360c1e8 docs: add subagent guidance and replay diff skill wuyang6 2026-06-12 14:39:07 +08:00
• c2f08821b9 sanitize admin tool usage stats wuyang6 2026-06-12 11:46:53 +08:00
• dbd279f333 improve admin usage analytics wuyang6 2026-06-12 11:44:48 +08:00
• 26e886fb57 allow fast update to restart system services wuyang6 2026-06-12 11:36:14 +08:00
• c4a0b6187b fix deploy service scope detection wuyang6 2026-06-12 11:32:58 +08:00
• 40d7c92a1d fix session routing and subagent visibility wuyang6 2026-06-12 11:06:04 +08:00
• a6e7d2260b 修改干预skill wangsenhao 2026-06-10 21:20:15 +08:00
• 54bc581cf2 修改label-master wangsenhao 2026-06-08 14:43:44 +08:00
• 84e792285b Update label master auto service labels wuyang6 2026-05-29 19:55:22 +08:00
• 4665b9a632 add lightweight model training skill wuyang6 2026-05-28 20:06:21 +08:00
• 98b3e586de fix: subprocess-based remote sync, pipeline status, skill updates hupenglong1 2026-05-28 14:35:35 +08:00
• 6364fb437c Normalize product data prompt formatting wuyang6 2026-05-28 14:30:55 +08:00
• 2dc941d6d8 Update vehicle control complexity rules wuyang6 2026-05-27 21:06:40 +08:00
• 1cefbf30c7 Merge branch 'wsh_dev' into 'main' 王森浩 2026-05-27 17:03:27 +08:00
• 85ce8c0e69 fix: session resume, config_set schema, metrics chart, subagent logging wangsenhao 2026-05-27 17:02:24 +08:00
• 74f55a421e fix: align model idle timeout with backend budget wuyang6 2026-05-26 17:21:24 +08:00
• 89e77140e7 chore(skill): slim down model-iteration skill files (-820 lines) wangsenhao 2026-05-26 17:03:08 +08:00
• 931cf2d065 fix: isolate auth cookies per deploy wuyang6 2026-05-26 16:24:11 +08:00
• 4dbf59116f docs: simplify complex_dev/complex_base, remove cleaned_predict row wangsenhao 2026-05-26 15:49:23 +08:00
• b48032e15c feat: backend gate fallback + metric_diff docs + submit_sft fix wangsenhao 2026-05-26 15:42:02 +08:00