dsh-vibe-math 2.7.3 → 2.8.0

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 (101) hide show
  1. package/LICENSE +21 -21
  2. package/README.en.md +682 -682
  3. package/README.md +683 -683
  4. package/docs/AUDIT-CHECKLIST.md +755 -744
  5. package/docs/COMPAT-AUDIT-ROUND2.md +351 -351
  6. package/docs/COMPAT-AUDIT-ROUND3.md +154 -154
  7. package/docs/audit-methodology.md +93 -0
  8. package/docs/final-paper.md +172 -172
  9. package/docs/formal-verification.md +613 -588
  10. package/docs/generate_framework_diagram_v2.py +114 -114
  11. package/docs/generate_framework_diagram_v3.py +127 -127
  12. package/docs/math-computation.md +195 -194
  13. package/docs/release-notes/README.en.md +1 -0
  14. package/docs/release-notes/README.md +1 -0
  15. package/docs/release-notes/RELEASE-NOTES-2.0.22.md +112 -112
  16. package/docs/release-notes/RELEASE-NOTES-2.1.0.md +143 -143
  17. package/docs/release-notes/RELEASE-NOTES-2.2.0.md +266 -266
  18. package/docs/release-notes/RELEASE-NOTES-2.2.1.md +43 -43
  19. package/docs/release-notes/RELEASE-NOTES-2.2.2.md +88 -88
  20. package/docs/release-notes/RELEASE-NOTES-2.3.0.md +207 -207
  21. package/docs/release-notes/RELEASE-NOTES-2.3.1.md +134 -134
  22. package/docs/release-notes/RELEASE-NOTES-2.3.10.md +105 -105
  23. package/docs/release-notes/RELEASE-NOTES-2.3.11.md +57 -57
  24. package/docs/release-notes/RELEASE-NOTES-2.3.12.md +80 -80
  25. package/docs/release-notes/RELEASE-NOTES-2.3.2.md +145 -145
  26. package/docs/release-notes/RELEASE-NOTES-2.3.3.md +115 -115
  27. package/docs/release-notes/RELEASE-NOTES-2.3.4.md +69 -69
  28. package/docs/release-notes/RELEASE-NOTES-2.3.5.md +63 -63
  29. package/docs/release-notes/RELEASE-NOTES-2.3.6.md +66 -66
  30. package/docs/release-notes/RELEASE-NOTES-2.3.7.md +59 -59
  31. package/docs/release-notes/RELEASE-NOTES-2.3.8.md +45 -45
  32. package/docs/release-notes/RELEASE-NOTES-2.3.9.md +70 -70
  33. package/docs/release-notes/RELEASE-NOTES-2.8.0.en.md +65 -0
  34. package/docs/release-notes/RELEASE-NOTES-2.8.0.md +65 -0
  35. package/docs/release-playbook.md +103 -0
  36. package/docs/slv-playbook.md +117 -0
  37. package/docs/test-timing.md +158 -157
  38. package/docs//346/236/266/346/236/204/345/233/276.md +77 -77
  39. package/installer.js +792 -792
  40. package/package.json +13 -3
  41. package/prompt-corpus-v2/formal-verify-v2.json +2 -2
  42. package/prompt-corpus-v2/formal-verify-v2.md +4 -0
  43. package/prompt-corpus-v3/formal-verify-v3.json +7 -7
  44. package/prompt-corpus-v3/formal-verify-v3.md +14 -0
  45. package/prompt-corpus-v5/prompt-corpus-v5.json +61 -61
  46. package/prompt-corpus-v5/prompt-corpus-v5.md +122 -0
  47. package/scripts/clean-temp.mjs +234 -0
  48. package/tests/audit-engine-faces.mjs +298 -0
  49. package/tests/audit-formal-sensitivity.mjs +345 -342
  50. package/tests/audit-installer-compat.test.mjs +256 -256
  51. package/tests/audit-math-computation-sensitivity.mjs +1 -1
  52. package/tests/audit-persona-sensitivity.mjs +323 -323
  53. package/tests/audit-persona-surface.test.mjs +431 -431
  54. package/tests/audit-prompt-duplication.mutants.mjs +3 -2
  55. package/tests/audit-prompt-invariants.mjs +589 -546
  56. package/tests/audit-registration.mjs +78 -78
  57. package/tests/audit-spec-traceability.mjs +193 -193
  58. package/tests/audit-status-report-fields.mjs +172 -172
  59. package/tests/audit-v5-integrity.mjs +795 -795
  60. package/tests/audit-v5-sensitivity.mjs +433 -432
  61. package/tests/e2e-v5-round2.test.mjs +2453 -2426
  62. package/tests/formal-verify-v2.test.mjs +1806 -1800
  63. package/tests/formal-verify-v3.test.mjs +1757 -1750
  64. package/tests/formal-verify-v4.test.mjs +2114 -2106
  65. package/tests/formal-verify-v5.test.mjs +774 -767
  66. package/tests/math-computation-discovery.mutants.mjs +3 -3
  67. package/tests/math-computation-shared.test.mjs +122 -5
  68. package/tests/math-computation-v2.test.mjs +14 -2
  69. package/tests/math-computation-v3.test.mjs +14 -2
  70. package/tests/math-computation-v4.test.mjs +739 -739
  71. package/tests/prompt-v5-integrity.test.mjs +1291 -1291
  72. package/tests/run-tests.mjs +393 -332
  73. package/tests/selfdrive-v5.mjs +560 -560
  74. package/tests/temp-hygiene.mutants.mjs +80 -0
  75. package/tests/v2-fix-probes.mutants.mjs +4 -0
  76. package/tests/v2-fix-probes.test.mjs +1034 -1033
  77. package/tests/v2v3-interrupt.mutants.mjs +111 -0
  78. package/tests/v3-fix-probes.mutants.mjs +99 -95
  79. package/tests/v3-fix-probes.test.mjs +891 -890
  80. package/tests/v4-final-paper.mutants.mjs +108 -0
  81. package/tests/v5-institute-fixes.mutants.mjs +218 -190
  82. package/vibe-math-v2/agent.cordis.yml +376 -376
  83. package/vibe-math-v2/math-computation.js +17 -15
  84. package/vibe-math-v2/preset.yml +2 -2
  85. package/vibe-math-v2/v2/347/232/204/351/242/235/345/244/226/350/246/201/346/261/202.txt +21 -21
  86. package/vibe-math-v2/vibe-math-v2.js +5078 -5011
  87. package/vibe-math-v2//345/256/236/347/216/260/346/226/271/346/241/210.md +486 -486
  88. package/vibe-math-v3/agent.cordis.yml +420 -420
  89. package/vibe-math-v3/math-computation.js +17 -15
  90. package/vibe-math-v3/preset.yml +2 -2
  91. package/vibe-math-v3/vibe-math-v3.js +6215 -6149
  92. package/vibe-math-v3//345/256/236/347/216/260/346/226/271/346/241/210.md +649 -649
  93. package/vibe-math-v4/agent.cordis.yml +438 -438
  94. package/vibe-math-v4/math-computation.js +17 -15
  95. package/vibe-math-v4/preset.yml +2 -2
  96. package/vibe-math-v4/vibe-math-v4.js +4490 -4422
  97. package/vibe-math-v4//345/256/236/347/216/260/346/226/271/346/241/210.md +1015 -1015
  98. package/vibe-math-v5/math-computation.js +17 -15
  99. package/vibe-math-v5/vibe-math-v5.js +93 -11
  100. package/vibe-math-v5//345/256/236/347/216/260/346/226/271/346/241/210.md +1728 -1728
  101. package/vibe-math-v5//346/236/266/346/236/204/345/233/276.md +693 -693
@@ -405,12 +405,12 @@ export async function probeMathEngines(host, opts) {
405
405
  if (name === 'cli') continue // cli needs a caller-supplied command, so it is not auto-detected
406
406
  const hit = await resolveCandidate(H, d)
407
407
  if (!hit) continue
408
- const v = await probeVersion(H, d, hit)
408
+ const v = await probeVersion(H, d, hit.exe)
409
409
  if (!v.ok) {
410
410
  if (d.license === 'commercial') continue // installed but not usable: reported by op=run as LICENSE_REQUIRED
411
411
  continue
412
412
  }
413
- engines.push(Object.assign({ name: name, version: v.version, path: hit, license: d.license }, verifyFields(name)))
413
+ engines.push(Object.assign({ name: name, version: v.version, path: hit.exe, foundVia: hit.via, license: d.license }, verifyFields(name)))
414
414
  }
415
415
  const result = { ok: true, engines: engines, available: engines.length > 0 }
416
416
  PROBE_CACHE.set(H, result)
@@ -461,7 +461,9 @@ async function knownInstallCandidates(H, engineName) {
461
461
  if (Array.isArray(r)) injected = r.filter((x) => typeof x === 'string' && x)
462
462
  } catch (e) { injected = null }
463
463
  }
464
- const roots = injected || mathInstallRoots(engineName, env, win)
464
+ // A JS empty array is TRUTHY: a host that answers `[]` must fall back to the built-in table,
465
+ // otherwise the whole per-OS stage is silently suppressed (task-2).
466
+ const roots = (Array.isArray(injected) && injected.length) ? injected : mathInstallRoots(engineName, env, win)
465
467
  const exts = win ? ['', '.exe', '.cmd'] : ['']
466
468
  const out = []
467
469
  const listNames = async (dir, kind) => {
@@ -491,17 +493,17 @@ async function resolveCandidate(H, d) {
491
493
  for (const c of cands) {
492
494
  try {
493
495
  const p = await H.resolveExecutable(c)
494
- if (typeof p === 'string' && p) return p
496
+ if (typeof p === 'string' && p) return { exe: p, via: 'path' }
495
497
  } catch (e) { /* not this one */ }
496
498
  }
497
499
  // round-7 (fix 2): DSH ships its own runtimes; PATH always wins, so this is a LAST resort. The
498
500
  // tree name is globbed (never hard-coded) and executable names come from the descriptor candidates.
499
501
  for (const p of await dshRuntimeCandidates(H, d.name)) {
500
- if (typeof p === 'string' && p) return p
502
+ if (typeof p === 'string' && p) return { exe: p, via: 'host-runtime' }
501
503
  }
502
504
  // round-9: then the known per-OS install locations (an engine installed without touching PATH).
503
505
  for (const p of await knownInstallCandidates(H, d.name)) {
504
- if (typeof p === 'string' && p) return p
506
+ if (typeof p === 'string' && p) return { exe: p, via: 'known-install' }
505
507
  }
506
508
  return null
507
509
  }
@@ -660,16 +662,16 @@ async function resolveEngine(H, requested, params, args) {
660
662
  cliProbe = fv.diag || null
661
663
  if (fv.ok) cliVersion = fv.version
662
664
  }
663
- return { ok: true, name: name, desc: d, exe: exe, version: cliVersion, family: fam || null, probe: cliProbe }
665
+ return { ok: true, name: name, desc: d, exe: exe, version: cliVersion, family: fam || null, probe: cliProbe, via: 'cli' }
664
666
  }
665
667
  const hit = await resolveCandidate(H, d)
666
668
  if (!hit) continue
667
- const v = await probeVersionWithRetry(H, d, hit)
669
+ const v = await probeVersionWithRetry(H, d, hit.exe)
668
670
  if (!v.ok) {
669
671
  if (d.license === 'commercial' || d.versionOptional) {
670
- const lic = await checkLicence(H, d, hit)
672
+ const lic = await checkLicence(H, d, hit.exe)
671
673
  if (!lic.ok) return fail('MATH_ENGINE_LICENSE_REQUIRED', name, name + ' 已安装但许可不可用(仅厂商可激活)', { next: next('vendor', Object.assign({ engine: name, url: d.vendor || '' }, verifyFields(name))) })
672
- return { ok: true, name: name, desc: d, exe: hit, version: 'unknown' }
674
+ return { ok: true, name: name, desc: d, exe: hit.exe, version: 'unknown', via: hit.via }
673
675
  }
674
676
  // round-7 (live-session fix): the engine WAS found - so an install guide would be misleading
675
677
  // (the user does not need to install anything; the host path is what failed). Report the probe
@@ -679,9 +681,9 @@ async function resolveEngine(H, requested, params, args) {
679
681
  next: next('note', { guidance: '引擎已找到但探测失败:请检查宿主 subprocess 通路(冷启动/超时/权限),或用 mathEngineOverride 调整 versionArgv 模板;已装的引擎不需要重装。' }),
680
682
  })
681
683
  }
682
- const lic = await checkLicence(H, d, hit)
684
+ const lic = await checkLicence(H, d, hit.exe)
683
685
  if (!lic.ok) return fail('MATH_ENGINE_LICENSE_REQUIRED', name, name + ' 已安装但许可不可用(仅厂商可激活)', { next: next('vendor', Object.assign({ engine: name, url: d.vendor || '' }, verifyFields(name))) })
684
- return { ok: true, name: name, desc: d, exe: hit, version: v.version }
686
+ return { ok: true, name: name, desc: d, exe: hit.exe, version: v.version, via: hit.via }
685
687
  }
686
688
  const first = requested === 'auto' ? (params.mathEngines[0] || 'python') : requested
687
689
  return fail('MATH_ENGINE_NOT_FOUND', first, '本机没有可用的计算引擎(试过:' + names.join(', ') + ')', { next: await userInstallNext(H, first) })
@@ -967,7 +969,7 @@ async function opProbe(H, args, params) {
967
969
  if (chosen && chosen.ok && !engines.some((e) => e.name === chosen.name)) {
968
970
  engines.push(Object.assign({ name: chosen.name, path: chosen.exe, version: chosen.version, license: (MATH_ENGINES[chosen.name] || {}).license }, verifyFields(chosen.name)))
969
971
  }
970
- const engineInfo = (chosen && chosen.ok) ? Object.assign({ name: chosen.name, path: chosen.exe, version: chosen.version }, verifyFields(chosen.name)) : null
972
+ const engineInfo = (chosen && chosen.ok) ? Object.assign({ name: chosen.name, path: chosen.exe, version: chosen.version, foundVia: chosen.via }, verifyFields(chosen.name)) : null
971
973
  // round-9 (F3): the availability line must never be silently partial - name the CONFIGURED engines
972
974
  // that were not found (and why), instead of just listing the ones that happened to resolve.
973
975
  const foundNames = engines.map((e) => e.name)
@@ -1000,7 +1002,7 @@ async function opRun(H, args, params) {
1000
1002
  const det = await resolveEngine(H, args.engine || 'auto', params, args)
1001
1003
  if (!det.ok) return det
1002
1004
  const desc = applyOverride(det, params, args)
1003
- const det2 = { ok: true, name: det.name, desc: desc, exe: det.exe, version: det.version }
1005
+ const det2 = { ok: true, name: det.name, desc: desc, exe: det.exe, version: det.version, via: det.via }
1004
1006
  // The generic escape hatch is recorded as `cli:<command>` (spec §4.3 / guards §18): the model and
1005
1007
  // the receipt must be able to tell WHICH command was run, not just that it was "cli".
1006
1008
  const cliCommand = args.cli && typeof args.cli.command === 'string' ? args.cli.command : ''
@@ -1118,7 +1120,7 @@ async function opRun(H, args, params) {
1118
1120
  schema: 'vibe-math/math-computation-receipt@1',
1119
1121
  runId: runId, preset: H.designator, project: slug(root), op: 'run', mode: mode,
1120
1122
  attempt: attempt, attemptDir: dir, baseRunDir: baseDir,
1121
- engine: { name: engineLabel, path: det.exe, version: det.version, source: det.name === 'cli' ? 'cli' : 'path' },
1123
+ engine: { name: engineLabel, path: det.exe, version: det.version, source: det.name === 'cli' ? 'cli' : 'path', foundVia: det2.via },
1122
1124
  script: { path: scriptRel, sha256: scriptHash, bytes: Buffer.byteLength(scriptText, 'utf8') },
1123
1125
  scriptPath: scriptRel, scriptHash: scriptHash,
1124
1126
  // round-8 (P0/D3): member file tools are session-cwd relative, so the ABSOLUTE locations must be
@@ -1,2 +1,2 @@
1
- name: Vibe Math V2
2
- description: 多代理数学问题求解与验证框架(经典):概率驱动调度 —— qs.json 问题清单 + Propos/ 命题库 + Reliable/ 参考文献,explorer 拆方向 → 逐方向 Solver 迭代 → 多验证器独立审查→辩论→裁决;断点续跑、人工/自动干预、file/push 双模式进度汇报。
1
+ name: Vibe Math V2
2
+ description: 多代理数学问题求解与验证框架(经典):概率驱动调度 —— qs.json 问题清单 + Propos/ 命题库 + Reliable/ 参考文献,explorer 拆方向 → 逐方向 Solver 迭代 → 多验证器独立审查→辩论→裁决;断点续跑、人工/自动干预、file/push 双模式进度汇报。
@@ -1,22 +1,22 @@
1
- 判断当前 v2 的下述是否已实现、当前是否符合下述,若未实现则请实现,过程中有不明确或觉得不合理或有自己建议时请询问我来决策:
2
-
3
- 1 .当前是否按我说的只允许“命题、命题+证明/证伪、问题+解法”被调度到验证器?以及是否按我说的只允许“问题”被调度到求解器solve。
4
-
5
- 2 .当前是否实现:当“命题”、“问题”的“证明/证伪/解法”的正确概率为1或0,或“命题”的布尔估计(正确概率)为1或0 或 问题已被解决时,则该证明/证伪/解法永不会被调度到去验证。当 “命题”、“问题”的证明/证伪/解法列表中存在它的正确概率为1的解法/证明/证伪时,则该“命题”、“问题”优先级设为 never,永不调度(到验证器、solve)。命题、问题 如果证明/证伪/解法列表非空且含有正确概率非0的证明/证伪/解法的话 ,则在验证时可优先验证“命题/问题+证明/证伪/解法”而不是“命题/问题”,在验证时优先调度其正确概率高的证明/证伪/解法进行验证。正确概率会动态调整。
6
-
7
- 3 .允许晋升为问题清单的“问题”的是“命题”,晋升原命题对象为“判断下述命题是否成立:...(原命题)”这样的问题对象,相应证明、证伪列表变为解法列表(并在各个证明证伪过程前面表明是要证明还是证伪),一些合理的、适配的属性也可以对应地转移过去。那些布尔概率为0、1的命题不会被晋升。
8
-
9
- 4 .那个Verified里需含有所有当前 被证明/证伪的命题和已解决的问题及它们的所有正确概率为1的证明/证伪/解法。
10
-
11
- 5 .按下述修改文件 实现方案.md 中的那个 q_sub 部分的逻辑(使得我的这个方案更严格化),然后按照下述来实现下述功能:q_sub 需为问题类对象,临时假设(记 p_{q-tmp} )为对 q_sub 的某种回答/答案的临时假设(是命题类对象)。把临时问题 q_sub 和“判断下述命题是否成立:p_{q-tmp}”交到问题清单(注意交到问题清单后的它们的概述、定义要完整,所涉及的一些依赖、对象等要给出完整定义,以避免断章取义),之后推理过程中发现的命题/结论中那些依赖这个假设的结论/命题在输出时需要把这个假设加上来作为命题/结论,如“若 p_{q-tmp} 成立,则...”(注意,概述、定义要完整,所依赖、基于的一些对象、环境、背景等要给出完整定义,不能断章取义)。
12
-
13
-
14
- 6 .是否有对代理通过提示词/上下文按照我实现方案中的定义来解释/告诉它各个参数的完整含义(如progress是指...)?
15
- 要完全对照我的实现方案,来对各提示词/上下文中可能导致代理理解不清楚的地方、对象、要求进行优化、完善、完整解释、举例解释等,优化提示词,使得能让ai/代理完全明确理解含义,使得符合我实现方案中的定义/含义/意思。一定要让代理完全明白、符合实现方案中的理解/要求来进行。
16
- 例如,一个命题、问题或其它各种对象及其各属性的完整定义(按照我实现方案.md中的来),只要1或0那么就视为绝对正确/错误,就可以被代理作为已知成立的结论来使用。
17
- 这样来保证它们在阅读文件时可以有正确的理解,以及保证它们可以明确到底要输出什么、对输出的内容有正确的理解、使得输出的内容符合要求(例如尤其是要求输出的那些对象一定要完整不要断章取义,要把环境/背景、一些依赖的对象/概念都完整补充、补充定义)。
18
- 当前是否有对它们介绍各个文件夹(里文件)的用处、各个文件夹里文件的格式及文件的各参数的含义、完整含义和用处(按我的实现方案来)。
19
-
20
- 7. 允许调控它们是否可以使用网络、写程序/脚本及其次数等。
21
-
1
+ 判断当前 v2 的下述是否已实现、当前是否符合下述,若未实现则请实现,过程中有不明确或觉得不合理或有自己建议时请询问我来决策:
2
+
3
+ 1 .当前是否按我说的只允许“命题、命题+证明/证伪、问题+解法”被调度到验证器?以及是否按我说的只允许“问题”被调度到求解器solve。
4
+
5
+ 2 .当前是否实现:当“命题”、“问题”的“证明/证伪/解法”的正确概率为1或0,或“命题”的布尔估计(正确概率)为1或0 或 问题已被解决时,则该证明/证伪/解法永不会被调度到去验证。当 “命题”、“问题”的证明/证伪/解法列表中存在它的正确概率为1的解法/证明/证伪时,则该“命题”、“问题”优先级设为 never,永不调度(到验证器、solve)。命题、问题 如果证明/证伪/解法列表非空且含有正确概率非0的证明/证伪/解法的话 ,则在验证时可优先验证“命题/问题+证明/证伪/解法”而不是“命题/问题”,在验证时优先调度其正确概率高的证明/证伪/解法进行验证。正确概率会动态调整。
6
+
7
+ 3 .允许晋升为问题清单的“问题”的是“命题”,晋升原命题对象为“判断下述命题是否成立:...(原命题)”这样的问题对象,相应证明、证伪列表变为解法列表(并在各个证明证伪过程前面表明是要证明还是证伪),一些合理的、适配的属性也可以对应地转移过去。那些布尔概率为0、1的命题不会被晋升。
8
+
9
+ 4 .那个Verified里需含有所有当前 被证明/证伪的命题和已解决的问题及它们的所有正确概率为1的证明/证伪/解法。
10
+
11
+ 5 .按下述修改文件 实现方案.md 中的那个 q_sub 部分的逻辑(使得我的这个方案更严格化),然后按照下述来实现下述功能:q_sub 需为问题类对象,临时假设(记 p_{q-tmp} )为对 q_sub 的某种回答/答案的临时假设(是命题类对象)。把临时问题 q_sub 和“判断下述命题是否成立:p_{q-tmp}”交到问题清单(注意交到问题清单后的它们的概述、定义要完整,所涉及的一些依赖、对象等要给出完整定义,以避免断章取义),之后推理过程中发现的命题/结论中那些依赖这个假设的结论/命题在输出时需要把这个假设加上来作为命题/结论,如“若 p_{q-tmp} 成立,则...”(注意,概述、定义要完整,所依赖、基于的一些对象、环境、背景等要给出完整定义,不能断章取义)。
12
+
13
+
14
+ 6 .是否有对代理通过提示词/上下文按照我实现方案中的定义来解释/告诉它各个参数的完整含义(如progress是指...)?
15
+ 要完全对照我的实现方案,来对各提示词/上下文中可能导致代理理解不清楚的地方、对象、要求进行优化、完善、完整解释、举例解释等,优化提示词,使得能让ai/代理完全明确理解含义,使得符合我实现方案中的定义/含义/意思。一定要让代理完全明白、符合实现方案中的理解/要求来进行。
16
+ 例如,一个命题、问题或其它各种对象及其各属性的完整定义(按照我实现方案.md中的来),只要1或0那么就视为绝对正确/错误,就可以被代理作为已知成立的结论来使用。
17
+ 这样来保证它们在阅读文件时可以有正确的理解,以及保证它们可以明确到底要输出什么、对输出的内容有正确的理解、使得输出的内容符合要求(例如尤其是要求输出的那些对象一定要完整不要断章取义,要把环境/背景、一些依赖的对象/概念都完整补充、补充定义)。
18
+ 当前是否有对它们介绍各个文件夹(里文件)的用处、各个文件夹里文件的格式及文件的各参数的含义、完整含义和用处(按我的实现方案来)。
19
+
20
+ 7. 允许调控它们是否可以使用网络、写程序/脚本及其次数等。
21
+
22
22
  8. 让对代理的提示词/上下文的一些关键部分(局部提示词/上下文)也作为 setting 值可配置的参数(方便用户自行修改/调控一些关键部分的提示词/上下文)。