@qwen-code/arena-agent 0.0.10-dev.10 → 0.0.10-dev.12

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (56) hide show
  1. package/.env.example +12 -1
  2. package/CHANGELOG.md +37 -1
  3. package/README.md +106 -28
  4. package/dist/agent.d.ts +1 -1
  5. package/dist/agent.js +90 -30
  6. package/dist/agent.js.map +1 -1
  7. package/dist/anthropic.js +17 -15
  8. package/dist/anthropic.js.map +1 -1
  9. package/dist/canonical-data.d.ts +2 -0
  10. package/dist/canonical-data.js +2 -0
  11. package/dist/canonical-data.js.map +1 -1
  12. package/dist/code-tools.js +34 -9
  13. package/dist/code-tools.js.map +1 -1
  14. package/dist/compaction.d.ts +3 -2
  15. package/dist/compaction.js +87 -21
  16. package/dist/compaction.js.map +1 -1
  17. package/dist/config.d.ts +6 -0
  18. package/dist/config.js +5 -0
  19. package/dist/config.js.map +1 -1
  20. package/dist/connector-mcp.d.ts +47 -0
  21. package/dist/connector-mcp.js +314 -0
  22. package/dist/connector-mcp.js.map +1 -0
  23. package/dist/connectors.d.ts +8 -0
  24. package/dist/connectors.js +57 -0
  25. package/dist/connectors.js.map +1 -0
  26. package/dist/midturn-checkpoint-hook.d.ts +14 -1
  27. package/dist/midturn-checkpoint-hook.js.map +1 -1
  28. package/dist/model-router-media.js +50 -0
  29. package/dist/model-router-media.js.map +1 -1
  30. package/dist/prompts.d.ts +0 -1
  31. package/dist/prompts.js +10 -36
  32. package/dist/prompts.js.map +1 -1
  33. package/dist/provider.js +3 -0
  34. package/dist/provider.js.map +1 -1
  35. package/dist/server.js +4 -2
  36. package/dist/server.js.map +1 -1
  37. package/dist/session-memory-prompts.js +1 -1
  38. package/dist/session-memory-prompts.js.map +1 -1
  39. package/dist/spec.d.ts +5 -4
  40. package/dist/spec.js +15 -9
  41. package/dist/spec.js.map +1 -1
  42. package/dist/{legacy-compaction.d.ts → summary-compaction.d.ts} +7 -0
  43. package/dist/{legacy-compaction.js → summary-compaction.js} +54 -10
  44. package/dist/summary-compaction.js.map +1 -0
  45. package/dist/tools.d.ts +3 -1
  46. package/dist/tools.js +35 -9
  47. package/dist/tools.js.map +1 -1
  48. package/dist/types.d.ts +4 -0
  49. package/dist/workspace.js +61 -14
  50. package/dist/workspace.js.map +1 -1
  51. package/docs/MODES_AND_RELEASES.md +55 -1
  52. package/modes/code-arena/full-stack/dependencies.lock.json +8670 -0
  53. package/modes/code-arena/react/dependencies.lock.json +2707 -0
  54. package/package.json +2 -1
  55. package/scripts/README.md +23 -4
  56. package/dist/legacy-compaction.js.map +0 -1
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@qwen-code/arena-agent",
3
- "version": "0.0.10-dev.10",
3
+ "version": "0.0.10-dev.12",
4
4
  "description": "A four-mode TypeScript agent harness for Arena and Code Arena traces.",
5
5
  "type": "module",
6
6
  "types": "./dist/sdk.d.ts",
@@ -73,6 +73,7 @@
73
73
  "typecheck": "tsc -p tsconfig.json --noEmit && tsc -p tsconfig.web.json --noEmit && npm run typecheck:code-arena-replay"
74
74
  },
75
75
  "dependencies": {
76
+ "@modelcontextprotocol/sdk": "1.32.1",
76
77
  "@stacksjs/ts-webp": "^0.1.2",
77
78
  "ajv": "^8.20.0",
78
79
  "ajv-formats": "^3.0.1",
package/scripts/README.md CHANGED
@@ -1,9 +1,28 @@
1
- # 轨迹复刻脚本
1
+ # 脚本入口
2
2
 
3
- 本目录保存 DataWorks SQL Arena 轨迹数据的复刻与环境准备脚本。数据集位于
4
- `tasks/`,运行产物位于 `tasks/reproductions/`。
3
+ 本目录包含构建、发布、对齐审计和历史实验脚本。先按用途选择入口,不能将
4
+ 旧实验脚本的默认样本或配置当作当前运行设置。
5
+
6
+ | 用途 | 入口 | 说明 |
7
+ | --- | --- | --- |
8
+ | 构建 CLI、Web 与 SDK | `npm run build` | 先生成当前契约和记忆提示词数据,再编译 TypeScript |
9
+ | 校验与发布二进制 | [发版流程](../docs/MODES_AND_RELEASES.md#发布节奏) | 包含版本、changelog 和平台包检查 |
10
+ | 更新 Code Arena 模板依赖 | `update-code-arena-dependency-locks.mjs --before <ISO date>` | 在空目录生成完整安装锁;更新后执行目标环境安装、构建验证 |
11
+ | 新批次 Agent Arena / Code Arena 对齐 | [arena-align](../skills/arena-align/SKILL.md) | 按用户指定输入组织四个阶段;不让模型重新决策 |
12
+ | Agent Arena 环境与契约审计 | `audit-arena-first-two.ts` | 当前阶段入口,保留不匹配样本;运行方式见对应 Skill |
13
+ | 旧 Agent Arena 审计 | `audit-arena-capture.ts` | 依赖旧 v2 工具基准,候选带精确契约过滤;不用于新批次完整审计 |
14
+ | 从旧 v2 数据重建契约 | `refresh-arena-v2-contract.mjs` | 显式覆盖当前 fixtures 的历史维护脚本;不是构建步骤,也不获取最新官方配置 |
15
+ | 旧任务的模型重跑实验 | `reproduce_case.mjs`、`reproduce_all.mjs` | 会调用 agent 模型重新执行任务,见下文;与历史工具回放的验证范围不同 |
16
+
17
+ `replay-*`、`audit-*`、`probe-*` 等工具中还有各批次的专项诊断入口。按关联的
18
+ Skill、测试和报告核对输入格式、执行身份及调用范围,不以文件是否出现在 npm
19
+ 命令列表里判断是否仍有用途。当前源码和生成文件的位置见
20
+ [源码入口与历史资料](../docs/MODES_AND_RELEASES.md#源码入口与历史资料)。
5
21
 
6
- ## 通用脚本
22
+ ## 旧任务的模型重跑实验
23
+
24
+ 以下说明仅适用于 DataWorks SQL Arena 数据的 `reproduce_*` 脚本。数据集位于
25
+ `tasks/`,运行产物位于 `tasks/reproductions/`。
7
26
 
8
27
  `reproduce_case.mjs` 运行单个 case,`reproduce_all.mjs`
9
28
  批量运行数据集中标记为可复刻的 case。
@@ -1 +0,0 @@
1
- {"version":3,"file":"legacy-compaction.js","sourceRoot":"","sources":["../src/legacy-compaction.ts"],"names":[],"mappings":"AAUA,MAAM,oBAAoB,GACxB,8EAA8E;IAC9E,+EAA+E;IAC/E,2EAA2E;IAC3E,6EAA6E;IAC7E,6EAA6E;IAC7E,yBAAyB,CAAC;AAE5B,MAAM,mBAAmB,GAAG;;;;;;;;;;;;;;;;4gBAgBgf,CAAC;AAE7gB,MAAM,cAAc,GAClB,8EAA8E,CAAC;AAEjF,MAAM,uBAAuB,GAC3B,+HAA+H,CAAC;AAoBlI,SAAS,iBAAiB,CAAC,OAAwB;IACjD,IAAI,OAAO,OAAO,CAAC,OAAO,KAAK,QAAQ;QAAE,OAAO,OAAO,CAAC,OAAO,CAAC;IAChE,OAAO,OAAO,CAAC,OAAO;SACnB,MAAM,CAAC,CAAC,IAAI,EAA4B,EAAE,CAAC,MAAM,IAAI,IAAI,CAAC;SAC1D,GAAG,CAAC,CAAC,IAAI,EAAE,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC;SACxB,IAAI,CAAC,EAAE,CAAC,CAAC;AACd,CAAC;AAED,SAAS,kBAAkB,CAAC,OAAgB;IAC1C,OAAO,CACL,OAAO,CAAC,IAAI,KAAK,MAAM;QACvB,KAAK,CAAC,OAAO,CAAC,OAAO,CAAC,OAAO,CAAC;QAC9B,uBAAuB,CAAC,IAAI,CAAC,iBAAiB,CAAC,OAAO,CAAC,CAAC,CACzD,CAAC;AACJ,CAAC;AAED,SAAS,oBAAoB,CAAC,QAA4B;IACxD,KAAK,IAAI,KAAK,GAAG,QAAQ,CAAC,MAAM,GAAG,CAAC,EAAE,KAAK,IAAI,CAAC,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QAC7D,MAAM,OAAO,GAAG,QAAQ,CAAC,KAAK,CAAC,CAAC;QAChC,IAAI,OAAO,EAAE,IAAI,KAAK,MAAM,IAAI,CAAC,kBAAkB,CAAC,OAAO,CAAC;YAAE,OAAO,KAAK,CAAC;IAC7E,CAAC;IACD,OAAO,CAAC,CAAC,CAAC;AACZ,CAAC;AAOD,SAAS,kBAAkB,CAAC,MAAuB,EAAE,KAAa;IAChE,KAAK,MAAM,SAAS,IAAI,KAAK,EAAE,CAAC;QAC9B,IAAI,SAAS,CAAC,WAAW,CAAC,CAAC,CAAE,IAAI,IAAI;YAAE,MAAM,CAAC,KAAK,IAAI,CAAC,CAAC;;YACpD,MAAM,CAAC,QAAQ,IAAI,CAAC,CAAC;IAC5B,CAAC;AACH,CAAC;AAED,SAAS,uBAAuB,CAC9B,QAA4B,EAC5B,KAAgC;IAEhC,MAAM,MAAM,GAAoB,EAAE,KAAK,EAAE,CAAC,EAAE,QAAQ,EAAE,CAAC,EAAE,CAAC;IAC1D,MAAM,QAAQ,GAAG,CAAC,GAAW,EAAE,KAAc,EAAW,EAAE,CACxD,GAAG,KAAK,OAAO,IAAI,OAAO,KAAK,KAAK,QAAQ,CAAC,CAAC,CAAC,SAAS,CAAC,CAAC,CAAC,KAAK,CAAC;IACnE,kBAAkB,CAAC,MAAM,EAAE,eAAe,CAAC,CAAC;IAC5C,KAAK,IAAI,KAAK,GAAG,CAAC,EAAE,KAAK,GAAG,QAAQ,CAAC,MAAM,EAAE,KAAK,IAAI,CAAC,EAAE,CAAC;QACxD,IAAI,KAAK;YAAE,kBAAkB,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC;QAC3C,MAAM,OAAO,GAAG,QAAQ,CAAC,KAAK,CAAE,CAAC;QACjC,MAAM,YAAY,GAChB,OAAO,CAAC,IAAI,KAAK,WAAW,IAAI,OAAO,CAAC,qBAAqB,EAAE,MAAM;YACnE,CAAC,CAAC;gBACE,IAAI,EAAE,OAAO,CAAC,IAAI;gBAClB,qBAAqB,EAAE,OAAO,CAAC,qBAAqB;aACrD;YACH,CAAC,CAAC,OAAO,CAAC;QACd,kBAAkB,CAAC,MAAM,EAAE,IAAI,CAAC,SAAS,CAAC,YAAY,EAAE,QAAQ,CAAC,CAAC,CAAC;IACrE,CAAC;IACD,kBAAkB,CAAC,MAAM,EAAE,YAAY,CAAC,CAAC;IACzC,kBAAkB,CAAC,MAAM,EAAE,IAAI,CAAC,SAAS,CAAC,KAAK,EAAE,QAAQ,CAAC,CAAC,CAAC;IAC5D,kBAAkB,CAAC,MAAM,EAAE,GAAG,CAAC,CAAC;IAChC,OAAO,IAAI,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,GAAG,CAAC,GAAG,MAAM,CAAC,QAAQ,CAAC,CAAC;AACvD,CAAC;AAED,SAAS,UAAU,CAAC,KAAc;IAChC,IAAI,KAAK,CAAC,OAAO,CAAC,KAAK,CAAC,EAAE,CAAC;QACzB,OAAO,KAAK,CAAC,MAAM,CAAC,CAAC,KAAK,EAAE,IAAI,EAAE,EAAE,CAAC,KAAK,GAAG,UAAU,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,CAAC;IACpE,CAAC;IACD,IAAI,CAAC,KAAK,IAAI,OAAO,KAAK,KAAK,QAAQ;QAAE,OAAO,CAAC,CAAC;IAClD,MAAM,MAAM,GAAG,KAAgC,CAAC;IAChD,MAAM,QAAQ,GACZ,OAAO,MAAM,CAAC,KAAK,KAAK,QAAQ,IAAI,MAAM,CAAC,IAAI,KAAK,WAAW,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAC1E,IAAI,YAAY,GAAG,CAAC,CAAC;IACrB,KAAK,MAAM,IAAI,IAAI,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC;QAAE,YAAY,IAAI,UAAU,CAAC,IAAI,CAAC,CAAC;IAC3E,OAAO,QAAQ,GAAG,YAAY,CAAC;AACjC,CAAC;AAED,MAAM,UAAU,qBAAqB,CACnC,QAA4B,EAC5B,KAAgC,EAChC,kBAAkB,GAAG,IAAI;IAEzB,MAAM,oBAAoB,GAAG,QAAQ,CAAC,GAAG,CAAC,CAAC,OAAO,EAAE,EAAE,CACpD,OAAO,CAAC,IAAI,KAAK,WAAW,IAAI,OAAO,CAAC,qBAAqB,EAAE,MAAM;QACnE,CAAC,CAAC;YACE,IAAI,EAAE,OAAO,CAAC,IAAI;YAClB,qBAAqB,EAAE,OAAO,CAAC,qBAAqB;SACrD;QACH,CAAC,CAAC,OAAO,CACZ,CAAC;IACF,MAAM,kBAAkB,GAAG,QAAQ,CAAC,MAAM,GAAG,CAAC,GAAG,KAAK,CAAC,MAAM,GAAG,EAAE,CAAC;IACnE,OAAO,CACL,uBAAuB,CAAC,QAAQ,EAAE,KAAK,CAAC;QACxC,UAAU,CAAC,oBAAoB,CAAC,GAAG,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,kBAAkB,CAAC;QAClE,kBAAkB,CACnB,CAAC;AACJ,CAAC;AAED,SAAS,WAAW,CAAC,KAAa;IAChC,OAAO,KAAK;SACT,IAAI,EAAE;SACN,OAAO,CAAC,0BAA0B,EAAE,8BAA8B,CAAC;SACnE,OAAO,CAAC,4BAA4B,EAAE,+BAA+B,CAAC,CAAC;AAC5E,CAAC;AAED,MAAM,UAAU,wBAAwB,CAAC,OAAe;IACtD,MAAM,WAAW,GAAG,WAAW,CAAC,OAAO,CAAC,CAAC,OAAO,CAAC,QAAQ,EAAE,EAAE,CAAC,CAAC;IAC/D,OAAO,CACL,0BAA0B;QAC1B,GAAG,oBAAoB,YAAY,WAAW,SAAS;QACvD,yBAAyB,CAC1B,CAAC;AACJ,CAAC;AAED,SAAS,cAAc,CACrB,OAAwB,EACxB,YAAoB;IAEpB,IAAI,OAAO,OAAO,CAAC,OAAO,KAAK,QAAQ,EAAE,CAAC;QACxC,OAAO;YACL,IAAI,EAAE,MAAM;YACZ,OAAO,EAAE,CAAC,EAAE,IAAI,EAAE,YAAY,EAAE,EAAE,EAAE,IAAI,EAAE,OAAO,CAAC,OAAO,EAAE,CAAC;SAC7B,CAAC;IACpC,CAAC;IACD,OAAO;QACL,IAAI,EAAE,MAAM;QACZ,OAAO,EAAE;YACP,EAAE,IAAI,EAAE,YAAY,EAAE;YACtB,GAAG,eAAe,CAAC,OAAO,CAAC,OAAO,CAAC;SACpC;KAC8B,CAAC;AACpC,CAAC;AAOD,KAAK,UAAU,eAAe,CAC5B,MAAmB,EACnB,KAAa,EACb,cAAkC,EAClC,SAAiB,EACjB,MAAoB;IAEpB,MAAM,OAAO,GAAiB;QAC5B,KAAK;QACL,QAAQ,EAAE;YACR,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,mBAAmB,EAAE;YAChD,GAAG,eAAe,CAAC,cAAc,CAAC;YAClC,EAAE,IAAI,EAAE,MAAM,EAAE,OAAO,EAAE,cAAc,EAAE;SAC1C;QACD,KAAK,EAAE,EAAE;QACT,UAAU,EAAE,MAAM;QAClB,WAAW,EAAE,CAAC;QACd,SAAS;QACT,cAAc,EAAE,KAAK;QACrB,cAAc,EAAE,CAAC;QACjB,gBAAgB,EAAE,KAAK;QACvB,MAAM;KACP,CAAC;IACF,MAAM,OAAO,GAAa,EAAE,CAAC;IAC7B,IAAI,KAA6B,CAAC;IAClC,IAAI,KAAK,EAAE,MAAM,KAAK,IAAI,MAAM,CAAC,MAAM,CAAC,OAAO,CAAC,EAAE,CAAC;QACjD,IAAI,KAAK,CAAC,KAAK;YAAE,KAAK,GAAG,KAAK,CAAC,KAAK,CAAC;QACrC,MAAM,KAAK,GAAG,KAAK,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,OAAO,CAAC;QACjD,IAAI,KAAK;YAAE,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC;IACzC,CAAC;IACD,OAAO,EAAE,OAAO,EAAE,WAAW,CAAC,OAAO,CAAC,IAAI,CAAC,EAAE,CAAC,CAAC,EAAE,KAAK,EAAE,CAAC;AAC3D,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,uBAAuB,CAC3C,QAAmB,EACnB,KAAgC,EAChC,MAAmB,EACnB,KAAa,EACb,QAA4B,EAC5B,MAAoB;IAEpB,MAAM,WAAW,GAAG,QAAQ,CAAC,kBAAkB,IAAI,IAAI,CAAC;IACxD,MAAM,MAAM,GAAG,qBAAqB,CAAC,QAAQ,EAAE,KAAK,EAAE,WAAW,CAAC,CAAC;IACnE,MAAM,SAAS,GAAG,QAAQ,CAAC,yBAAyB,IAAI,CACtD,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,QAAQ,CAAC,aAAa,CAAC;QACnC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,QAAQ,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC;IACnD,IAAI,MAAM,GAAG,SAAS,EAAE,CAAC;QACvB,OAAO;YACL,SAAS,EAAE,KAAK;YAChB,qBAAqB,EAAE,MAAM;YAC7B,oBAAoB,EAAE,MAAM;SAC7B,CAAC;IACJ,CAAC;IAED,MAAM,gBAAgB,GAAG,oBAAoB,CAAC,QAAQ,CAAC,CAAC;IACxD,IAAI,gBAAgB,IAAI,CAAC,EAAE,CAAC;QAC1B,OAAO;YACL,SAAS,EAAE,KAAK;YAChB,qBAAqB,EAAE,MAAM;YAC7B,oBAAoB,EAAE,MAAM;SAC7B,CAAC;IACJ,CAAC;IAED,MAAM,EAAE,OAAO,EAAE,KAAK,EAAE,GAAG,MAAM,eAAe,CAC9C,MAAM,EACN,KAAK,EACL,QAAQ,CAAC,KAAK,CAAC,CAAC,EAAE,gBAAgB,CAAC,EACnC,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,gBAAgB,CAAC,CAAC,EAClD,MAAM,CACP,CAAC;IACF,IAAI,CAAC,OAAO,EAAE,CAAC;QACb,OAAO;YACL,SAAS,EAAE,KAAK;YAChB,qBAAqB,EAAE,MAAM;YAC7B,oBAAoB,EAAE,MAAM;YAC5B,KAAK;SACN,CAAC;IACJ,CAAC;IAED,MAAM,WAAW,GAAG,QAAQ,CAAC,gBAAgB,CAAC,CAAC;IAC/C,IAAI,WAAW,EAAE,IAAI,KAAK,MAAM,EAAE,CAAC;QACjC,OAAO;YACL,SAAS,EAAE,KAAK;YAChB,qBAAqB,EAAE,MAAM;YAC7B,oBAAoB,EAAE,MAAM;YAC5B,KAAK;SACN,CAAC;IACJ,CAAC;IACD,MAAM,iBAAiB,GAAc;QACnC,QAAQ,CAAC,CAAC,CAAE;QACZ,cAAc,CAAC,WAAW,EAAE,wBAAwB,CAAC,OAAO,CAAC,CAAC;QAC9D,GAAG,QAAQ,CAAC,KAAK,CAAC,gBAAgB,GAAG,CAAC,CAAC;KACxC,CAAC;IACF,QAAQ,CAAC,MAAM,CAAC,CAAC,EAAE,QAAQ,CAAC,MAAM,EAAE,GAAG,iBAAiB,CAAC,CAAC;IAC1D,OAAO;QACL,SAAS,EAAE,IAAI;QACf,qBAAqB,EAAE,MAAM;QAC7B,oBAAoB,EAAE,qBAAqB,CAAC,QAAQ,EAAE,KAAK,EAAE,WAAW,CAAC;QACzE,KAAK;QACL,MAAM,EAAE,OAAO;QACf,eAAe,EAAE,gBAAgB;KAClC,CAAC;AACJ,CAAC"}