oh-my-knowledge 0.26.0 → 0.28.0

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 (167) hide show
  1. package/README.md +105 -346
  2. package/README.zh.md +142 -375
  3. package/dist/src/analysis/report-diagnostics.d.ts +2 -2
  4. package/dist/src/analysis/report-diagnostics.js +2 -2
  5. package/dist/src/analysis/sample-diagnostics.d.ts +3 -3
  6. package/dist/src/analysis/sample-diagnostics.js +3 -3
  7. package/dist/src/authoring/generator.d.ts.map +1 -1
  8. package/dist/src/authoring/generator.js +9 -8
  9. package/dist/src/authoring/generator.js.map +1 -1
  10. package/dist/src/cli/cli-exit.d.ts +15 -0
  11. package/dist/src/cli/cli-exit.d.ts.map +1 -0
  12. package/dist/src/cli/cli-exit.js +19 -0
  13. package/dist/src/cli/cli-exit.js.map +1 -0
  14. package/dist/src/cli/commands/_shared.d.ts +12 -0
  15. package/dist/src/cli/commands/_shared.d.ts.map +1 -0
  16. package/dist/src/cli/commands/_shared.js +26 -0
  17. package/dist/src/cli/commands/_shared.js.map +1 -0
  18. package/dist/src/cli/commands/doctor.d.ts +2 -0
  19. package/dist/src/cli/commands/doctor.d.ts.map +1 -0
  20. package/dist/src/cli/commands/doctor.js +174 -0
  21. package/dist/src/cli/commands/doctor.js.map +1 -0
  22. package/dist/src/cli/commands/eval-gold.d.ts +2 -0
  23. package/dist/src/cli/commands/eval-gold.d.ts.map +1 -0
  24. package/dist/src/cli/commands/eval-gold.js +137 -0
  25. package/dist/src/cli/commands/eval-gold.js.map +1 -0
  26. package/dist/src/cli/commands/eval-runner.d.ts +2 -0
  27. package/dist/src/cli/commands/eval-runner.d.ts.map +1 -0
  28. package/dist/src/cli/commands/eval-runner.js +299 -0
  29. package/dist/src/cli/commands/eval-runner.js.map +1 -0
  30. package/dist/src/cli/commands/eval.d.ts +2 -0
  31. package/dist/src/cli/commands/eval.d.ts.map +1 -0
  32. package/dist/src/cli/commands/eval.js +11 -0
  33. package/dist/src/cli/commands/eval.js.map +1 -0
  34. package/dist/src/cli/commands/evolve.d.ts +2 -0
  35. package/dist/src/cli/commands/evolve.d.ts.map +1 -0
  36. package/dist/src/cli/commands/evolve.js +115 -0
  37. package/dist/src/cli/commands/evolve.js.map +1 -0
  38. package/dist/src/cli/commands/init.d.ts +2 -0
  39. package/dist/src/cli/commands/init.d.ts.map +1 -0
  40. package/dist/src/cli/commands/init.js +110 -0
  41. package/dist/src/cli/commands/init.js.map +1 -0
  42. package/dist/src/cli/commands/observe.d.ts +2 -0
  43. package/dist/src/cli/commands/observe.d.ts.map +1 -0
  44. package/dist/src/cli/commands/observe.js +77 -0
  45. package/dist/src/cli/commands/observe.js.map +1 -0
  46. package/dist/src/cli/commands/registry.d.ts +11 -0
  47. package/dist/src/cli/commands/registry.d.ts.map +1 -0
  48. package/dist/src/cli/commands/registry.js +24 -0
  49. package/dist/src/cli/commands/registry.js.map +1 -0
  50. package/dist/src/cli/commands/sample.d.ts +2 -0
  51. package/dist/src/cli/commands/sample.d.ts.map +1 -0
  52. package/dist/src/cli/commands/sample.js +121 -0
  53. package/dist/src/cli/commands/sample.js.map +1 -0
  54. package/dist/src/cli/commands/studio.d.ts +2 -0
  55. package/dist/src/cli/commands/studio.d.ts.map +1 -0
  56. package/dist/src/cli/commands/studio.js +76 -0
  57. package/dist/src/cli/commands/studio.js.map +1 -0
  58. package/dist/src/cli/i18n-dict.d.ts +2 -4
  59. package/dist/src/cli/i18n-dict.d.ts.map +1 -1
  60. package/dist/src/cli/i18n-dict.js +461 -675
  61. package/dist/src/cli/i18n-dict.js.map +1 -1
  62. package/dist/src/cli/index.js +43 -1516
  63. package/dist/src/cli/index.js.map +1 -1
  64. package/dist/src/cli/parse-run-config.d.ts +3 -4
  65. package/dist/src/cli/parse-run-config.d.ts.map +1 -1
  66. package/dist/src/cli/parse-run-config.js +5 -5
  67. package/dist/src/cli/parse-run-config.js.map +1 -1
  68. package/dist/src/cli/parse-strict.d.ts +0 -13
  69. package/dist/src/cli/parse-strict.d.ts.map +1 -1
  70. package/dist/src/cli/parse-strict.js +2 -1
  71. package/dist/src/cli/parse-strict.js.map +1 -1
  72. package/dist/src/doctor/health/builtin-dimensions.d.ts +11 -0
  73. package/dist/src/doctor/health/builtin-dimensions.d.ts.map +1 -0
  74. package/dist/src/doctor/health/builtin-dimensions.js +94 -0
  75. package/dist/src/doctor/health/builtin-dimensions.js.map +1 -0
  76. package/dist/src/doctor/health/composer.d.ts +19 -0
  77. package/dist/src/doctor/health/composer.d.ts.map +1 -0
  78. package/dist/src/doctor/health/composer.js +289 -0
  79. package/dist/src/doctor/health/composer.js.map +1 -0
  80. package/dist/src/doctor/health/dimension-registry.d.ts +13 -0
  81. package/dist/src/doctor/health/dimension-registry.d.ts.map +1 -0
  82. package/dist/src/doctor/health/dimension-registry.js +28 -0
  83. package/dist/src/doctor/health/dimension-registry.js.map +1 -0
  84. package/dist/src/doctor/health/dimension-spec.d.ts +46 -0
  85. package/dist/src/doctor/health/dimension-spec.d.ts.map +1 -0
  86. package/dist/src/doctor/health/dimension-spec.js +12 -0
  87. package/dist/src/doctor/health/dimension-spec.js.map +1 -0
  88. package/dist/src/doctor/health/parser.d.ts +27 -0
  89. package/dist/src/doctor/health/parser.d.ts.map +1 -0
  90. package/dist/src/doctor/health/parser.js +190 -0
  91. package/dist/src/doctor/health/parser.js.map +1 -0
  92. package/dist/src/doctor/health/prompt-builder.d.ts +22 -0
  93. package/dist/src/doctor/health/prompt-builder.d.ts.map +1 -0
  94. package/dist/src/doctor/health/prompt-builder.js +162 -0
  95. package/dist/src/doctor/health/prompt-builder.js.map +1 -0
  96. package/dist/src/doctor/health/register.d.ts +13 -0
  97. package/dist/src/doctor/health/register.d.ts.map +1 -0
  98. package/dist/src/doctor/health/register.js +20 -0
  99. package/dist/src/doctor/health/register.js.map +1 -0
  100. package/dist/src/doctor/html-renderer.d.ts +20 -0
  101. package/dist/src/doctor/html-renderer.d.ts.map +1 -0
  102. package/dist/src/doctor/html-renderer.js +366 -0
  103. package/dist/src/doctor/html-renderer.js.map +1 -0
  104. package/dist/src/doctor/index.d.ts +1 -1
  105. package/dist/src/doctor/index.d.ts.map +1 -1
  106. package/dist/src/doctor/index.js +57 -8
  107. package/dist/src/doctor/index.js.map +1 -1
  108. package/dist/src/doctor/preflight.d.ts +2 -2
  109. package/dist/src/doctor/preflight.js +2 -2
  110. package/dist/src/doctor/renderer.d.ts +4 -0
  111. package/dist/src/doctor/renderer.d.ts.map +1 -1
  112. package/dist/src/doctor/renderer.js +72 -18
  113. package/dist/src/doctor/renderer.js.map +1 -1
  114. package/dist/src/doctor/rules.d.ts +6 -5
  115. package/dist/src/doctor/rules.d.ts.map +1 -1
  116. package/dist/src/doctor/rules.js +4 -3
  117. package/dist/src/doctor/rules.js.map +1 -1
  118. package/dist/src/eval-core/fact-checker.js +1 -1
  119. package/dist/src/eval-core/fact-checker.js.map +1 -1
  120. package/dist/src/eval-core/layer-gates.d.ts +1 -1
  121. package/dist/src/eval-core/layer-gates.js +1 -1
  122. package/dist/src/eval-core/verdict.d.ts +4 -4
  123. package/dist/src/eval-core/verdict.d.ts.map +1 -1
  124. package/dist/src/eval-core/verdict.js +2 -2
  125. package/dist/src/eval-workflows/batch-evaluation-workflow.d.ts +15 -2
  126. package/dist/src/eval-workflows/batch-evaluation-workflow.d.ts.map +1 -1
  127. package/dist/src/eval-workflows/batch-evaluation-workflow.js +13 -2
  128. package/dist/src/eval-workflows/batch-evaluation-workflow.js.map +1 -1
  129. package/dist/src/eval-workflows/evaluation-pipeline.d.ts +6 -3
  130. package/dist/src/eval-workflows/evaluation-pipeline.d.ts.map +1 -1
  131. package/dist/src/eval-workflows/evaluation-pipeline.js +10 -9
  132. package/dist/src/eval-workflows/evaluation-pipeline.js.map +1 -1
  133. package/dist/src/eval-workflows/run-evaluation.d.ts +1 -1
  134. package/dist/src/eval-workflows/run-evaluation.d.ts.map +1 -1
  135. package/dist/src/eval-workflows/run-evaluation.js +13 -6
  136. package/dist/src/eval-workflows/run-evaluation.js.map +1 -1
  137. package/dist/src/executors/script.d.ts.map +1 -1
  138. package/dist/src/executors/script.js +16 -3
  139. package/dist/src/executors/script.js.map +1 -1
  140. package/dist/src/grading/debias-validate.d.ts +2 -2
  141. package/dist/src/grading/debias-validate.js +2 -2
  142. package/dist/src/grading/gold-cli.d.ts +2 -5
  143. package/dist/src/grading/gold-cli.d.ts.map +1 -1
  144. package/dist/src/grading/gold-cli.js +4 -8
  145. package/dist/src/grading/gold-cli.js.map +1 -1
  146. package/dist/src/grading/judge.d.ts +1 -1
  147. package/dist/src/renderer/html-renderer.js +1 -1
  148. package/dist/src/renderer/layout.js +5 -5
  149. package/dist/src/renderer/layout.js.map +1 -1
  150. package/dist/src/renderer/skill-health-renderer.d.ts +1 -1
  151. package/dist/src/renderer/skill-health-renderer.js +1 -1
  152. package/dist/src/renderer/summary.js +8 -8
  153. package/dist/src/server/report-server.d.ts.map +1 -1
  154. package/dist/src/server/report-server.js +8 -7
  155. package/dist/src/server/report-server.js.map +1 -1
  156. package/dist/src/types/doctor.d.ts +44 -6
  157. package/dist/src/types/doctor.d.ts.map +1 -1
  158. package/dist/src/types/doctor.js +3 -0
  159. package/dist/src/types/doctor.js.map +1 -1
  160. package/dist/src/types/eval.d.ts +1 -1
  161. package/dist/src/types/report.d.ts +2 -2
  162. package/dist/src/types/report.d.ts.map +1 -1
  163. package/package.json +1 -1
  164. package/dist/src/cli/coverage-renderer.d.ts +0 -15
  165. package/dist/src/cli/coverage-renderer.d.ts.map +0 -1
  166. package/dist/src/cli/coverage-renderer.js +0 -74
  167. package/dist/src/cli/coverage-renderer.js.map +0 -1
@@ -0,0 +1,299 @@
1
+ import { CliExit } from '../cli-exit.js';
2
+ import { tCli, langFromArgv } from '../i18n.js';
3
+ import { parseRunConfig } from '../parse-run-config.js';
4
+ import { makeOnProgress } from '../progress.js';
5
+ import { computeRunTally } from '../run-tally.js';
6
+ function isDryRunReport(report) {
7
+ return Boolean(report && typeof report === 'object' && report.dryRun === true);
8
+ }
9
+ function isDryRunBatchReport(report) {
10
+ return isDryRunReport(report) && report.batch === true;
11
+ }
12
+ function verdictOptions(values) {
13
+ const rawThreshold = values.threshold;
14
+ const gateThreshold = rawThreshold !== undefined && Number.isFinite(Number(rawThreshold))
15
+ ? Number(rawThreshold)
16
+ : 3.5;
17
+ const rawTrivial = values['trivial-diff'];
18
+ const triviallySmallDiff = rawTrivial !== undefined && Number.isFinite(Number(rawTrivial))
19
+ ? Number(rawTrivial)
20
+ : undefined;
21
+ return { gateThreshold, triviallySmallDiff };
22
+ }
23
+ function verdictPasses(level, headline) {
24
+ return level === 'PROGRESS' || (level === 'SOLO' && headline.includes('PASS'));
25
+ }
26
+ function reportOnlyMode(values) {
27
+ return values['report-only'] === true || values['no-gate'] === true;
28
+ }
29
+ function applyGateExitCode(code, values, lang) {
30
+ if (!reportOnlyMode(values))
31
+ return code;
32
+ process.stderr.write(tCli('cli.run.report_only_gate_skipped', lang));
33
+ return 0;
34
+ }
35
+ async function emitEvaluationVerdict(report, values) {
36
+ const { computeVerdict, formatVerdictText } = await import('../../eval-core/verdict.js');
37
+ const result = computeVerdict(report, verdictOptions(values));
38
+ console.log(formatVerdictText(result, { verbose: true }));
39
+ return verdictPasses(result.level, result.headline) ? 0 : 1;
40
+ }
41
+ function batchItemFallbackReport(batch, item) {
42
+ return {
43
+ kind: 'evaluation',
44
+ id: item.reportId,
45
+ meta: {
46
+ ...batch.meta,
47
+ variants: ['baseline', item.name],
48
+ sampleCount: item.sampleCount,
49
+ totalCostUSD: item.totalCostUSD,
50
+ artifactHashes: item.artifactHash ? { [item.name]: item.artifactHash } : {},
51
+ },
52
+ summary: item.summary,
53
+ results: [],
54
+ ...(item.variance ? { variance: item.variance } : {}),
55
+ };
56
+ }
57
+ async function loadBatchChildReports(batch, reportsDir, lang) {
58
+ const { createFileStore } = await import('../../server/report-store.js');
59
+ const store = createFileStore(reportsDir);
60
+ const reports = [];
61
+ for (const item of batch.items) {
62
+ const loaded = await store.get(item.reportId);
63
+ if (loaded?.kind === 'evaluation') {
64
+ reports.push(loaded);
65
+ }
66
+ else {
67
+ process.stderr.write(tCli('cli.run.batch_child_report_missing', lang, { id: item.reportId }));
68
+ reports.push(batchItemFallbackReport(batch, item));
69
+ }
70
+ }
71
+ return reports;
72
+ }
73
+ async function emitBatchVerdict(report, reportsDir, values, lang) {
74
+ const { computeVerdict } = await import('../../eval-core/verdict.js');
75
+ const childReports = await loadBatchChildReports(report, reportsDir, lang);
76
+ const results = childReports.map((child) => ({
77
+ id: child.id,
78
+ treatment: child.meta.variants[1] ?? child.id,
79
+ verdict: computeVerdict(child, verdictOptions(values)),
80
+ }));
81
+ const passed = results.filter((r) => verdictPasses(r.verdict.level, r.verdict.headline)).length;
82
+ const failed = results.length - passed;
83
+ const status = lang === 'zh'
84
+ ? (failed === 0 ? '通过' : '未通过')
85
+ : (failed === 0 ? 'PASS' : 'FAIL');
86
+ console.log(tCli('cli.run.batch_verdict_header', lang, {
87
+ status,
88
+ passed,
89
+ total: results.length,
90
+ }));
91
+ for (const result of results) {
92
+ console.log(` ${result.verdict.level}: ${result.treatment} — ${result.verdict.headline}`);
93
+ }
94
+ return failed === 0 ? 0 : 1;
95
+ }
96
+ async function announceSavedReport({ report, filePath, reportsDir, values, lang, }) {
97
+ const tally = computeRunTally(report);
98
+ process.stderr.write(tCli(report.kind === 'batch-evaluation' ? 'cli.run.batch_complete' : 'cli.run.eval_complete', lang));
99
+ process.stderr.write(tCli('cli.run.tally', lang, tally));
100
+ process.stderr.write(tCli('cli.run.report_saved', lang, { path: filePath }));
101
+ if (!values['no-serve'] && process.stdout.isTTY) {
102
+ const { createReportServer } = await import('../../server/report-server.js');
103
+ const server = createReportServer({ reportsDir });
104
+ const serverUrl = await server.start();
105
+ const reportUrl = `${serverUrl}/reports/${report.id}`;
106
+ process.stderr.write(tCli('cli.run.report_server_running', lang, { url: serverUrl }));
107
+ process.stderr.write(tCli('cli.run.report_server_view', lang, { url: reportUrl }));
108
+ process.stderr.write(tCli('cli.run.report_server_stop', lang));
109
+ const { platform } = await import('node:os');
110
+ const openCmd = platform() === 'darwin' ? 'open' : platform() === 'win32' ? 'start' : 'xdg-open';
111
+ const { execFile: execFileCb } = await import('node:child_process');
112
+ execFileCb(openCmd, [reportUrl], () => undefined);
113
+ }
114
+ else if (!values['no-serve']) {
115
+ process.stderr.write(tCli('cli.run.no_serve_in_non_tty', lang));
116
+ process.stderr.write(tCli('cli.run.no_serve_view_hint', lang, { id: report.id, dir: reportsDir }));
117
+ }
118
+ }
119
+ export async function execute(argv) {
120
+ const lang = langFromArgv(argv);
121
+ // 注: 这里**不**给 parseArgs default 值, 否则 values.xxx 永远不为 undefined,
122
+ // CLI > eval.yaml > hardcoded-default 三级 fallback 区分不开 ("用户没传" vs "用户传了等于 default 值")。
123
+ // hardcoded default 在下面处理 undefined 时显式给。
124
+ const { values, config, evalConfig } = parseRunConfig(argv, {
125
+ blind: { type: 'boolean' },
126
+ repeat: { type: 'string' },
127
+ 'judge-repeat': { type: 'string' },
128
+ bootstrap: { type: 'boolean' },
129
+ 'bootstrap-samples': { type: 'string' },
130
+ 'gold-dir': { type: 'string' },
131
+ 'no-debias-length': { type: 'boolean' },
132
+ 'budget-usd': { type: 'string' },
133
+ 'budget-per-sample-usd': { type: 'string' },
134
+ 'budget-per-sample-ms': { type: 'string' },
135
+ threshold: { type: 'string', default: '3.5' },
136
+ 'trivial-diff': { type: 'string' },
137
+ 'report-only': { type: 'boolean' },
138
+ 'no-gate': { type: 'boolean' },
139
+ });
140
+ const { runEvaluation, runMultiple, runBatchEvaluation } = await import('../../eval-workflows/run-evaluation.js');
141
+ if (values.blind !== undefined) {
142
+ config.blind = values.blind;
143
+ }
144
+ config.onProgress = makeOnProgress(lang);
145
+ // --repeat: CLI > eval.yaml > 1. 非 ≥1 整数时提示并钳到 1。
146
+ const repeatRaw = values.repeat;
147
+ const parsedRepeat = repeatRaw !== undefined ? Number(repeatRaw) : (evalConfig?.repeat ?? 1);
148
+ if (repeatRaw !== undefined && (!Number.isFinite(parsedRepeat) || parsedRepeat < 1)) {
149
+ process.stderr.write(tCli('cli.run.invalid_repeat', lang, { value: repeatRaw }));
150
+ }
151
+ const repeatCount = Math.max(1, Math.floor(parsedRepeat) || 1);
152
+ // --judge-repeat: CLI > eval.yaml > 1.
153
+ const judgeRepeatRaw = values['judge-repeat'];
154
+ const parsedJudgeRepeat = judgeRepeatRaw !== undefined ? Number(judgeRepeatRaw) : (evalConfig?.judgeRepeat ?? 1);
155
+ if (judgeRepeatRaw !== undefined && (!Number.isFinite(parsedJudgeRepeat) || parsedJudgeRepeat < 1)) {
156
+ process.stderr.write(tCli('cli.run.invalid_judge_repeat', lang, { value: judgeRepeatRaw }));
157
+ }
158
+ const judgeRepeatCount = Math.max(1, Math.floor(parsedJudgeRepeat) || 1);
159
+ if (judgeRepeatCount > 1)
160
+ config.judgeRepeat = judgeRepeatCount;
161
+ // --budget-usd / --budget-per-sample-usd / --budget-per-sample-ms:
162
+ // hard budget caps. CLI flags override config-file values. When the
163
+ // total-USD cap is exceeded mid-run, remaining tasks are skipped and a
164
+ // partial report is persisted with meta.budgetExhausted=true.
165
+ const budgetUSD = values['budget-usd'] != null ? Number(values['budget-usd']) : undefined;
166
+ const budgetPerSampleUSD = values['budget-per-sample-usd'] != null ? Number(values['budget-per-sample-usd']) : undefined;
167
+ const budgetPerSampleMs = values['budget-per-sample-ms'] != null ? Number(values['budget-per-sample-ms']) : undefined;
168
+ if (budgetUSD !== undefined || budgetPerSampleUSD !== undefined || budgetPerSampleMs !== undefined) {
169
+ config.budget = {
170
+ ...(budgetUSD !== undefined && Number.isFinite(budgetUSD) && budgetUSD >= 0 ? { totalUSD: budgetUSD } : {}),
171
+ ...(budgetPerSampleUSD !== undefined && Number.isFinite(budgetPerSampleUSD) && budgetPerSampleUSD >= 0 ? { perSampleUSD: budgetPerSampleUSD } : {}),
172
+ ...(budgetPerSampleMs !== undefined && Number.isFinite(budgetPerSampleMs) && budgetPerSampleMs >= 0 ? { perSampleMs: budgetPerSampleMs } : {}),
173
+ };
174
+ }
175
+ // --no-debias-length / eval.yaml `lengthDebias: false`: opt out of length-controlled prompt。
176
+ // Default debias-on (judge prompt v3-cot-length); flip off only to reproduce historical reports。
177
+ // CLI 显式 --no-debias-length > eval.yaml lengthDebias > 默认 true。
178
+ const lengthDebiasOff = values['no-debias-length'] === true
179
+ || (values['no-debias-length'] === undefined && evalConfig?.lengthDebias === false);
180
+ if (lengthDebiasOff) {
181
+ config.lengthDebias = false;
182
+ process.stderr.write(tCli('cli.run.no_debias_length_active', lang));
183
+ }
184
+ // --bootstrap / --bootstrap-samples: CLI > eval.yaml > default(on / 1000)。
185
+ // `omk eval` owns the product default here instead of wrapper-injecting argv
186
+ // in commands/eval.ts, so config-file opt-out stays possible.
187
+ const bootstrapEnabled = values.bootstrap === true
188
+ || (values.bootstrap === undefined && evalConfig?.bootstrap !== false);
189
+ if (bootstrapEnabled) {
190
+ config.bootstrap = true;
191
+ const bsRaw = values['bootstrap-samples'];
192
+ const parsedBs = bsRaw !== undefined ? Number(bsRaw) : (evalConfig?.bootstrapSamples ?? 1000);
193
+ if (bsRaw !== undefined && (!Number.isFinite(parsedBs) || parsedBs < 100)) {
194
+ process.stderr.write(tCli('cli.run.invalid_bootstrap_samples', lang, { value: bsRaw }));
195
+ }
196
+ const bsCount = Math.max(100, Math.floor(parsedBs) || 1000);
197
+ if (bsCount > 10000) {
198
+ process.stderr.write(tCli('cli.run.bootstrap_samples_too_large', lang, { n: bsCount }));
199
+ }
200
+ config.bootstrapSamples = bsCount;
201
+ }
202
+ // 注入 lang 让 evaluation pipeline 能渲染 doctor 报告(失败时)。
203
+ config.lang = lang;
204
+ if (values['skip-connectivity']) {
205
+ process.stderr.write(tCli('cli.run.skip_connectivity_warning', lang) + '\n');
206
+ }
207
+ try {
208
+ // --batch mode: evaluate each skill independently
209
+ if (values.batch) {
210
+ const { report, filePath } = await runBatchEvaluation({
211
+ ...config,
212
+ repeat: repeatCount,
213
+ onSkillProgress({ phase, skill, current, total }) {
214
+ if (phase === 'start') {
215
+ process.stderr.write(tCli('cli.run.skill_section', lang, {
216
+ i: current ?? '', n: total ?? '', skill: skill ?? '',
217
+ }));
218
+ }
219
+ },
220
+ });
221
+ console.log(JSON.stringify(report, null, 2));
222
+ if (isDryRunBatchReport(report)) {
223
+ console.log(tCli('cli.run.dry_run_no_scores', lang));
224
+ throw new CliExit(0);
225
+ }
226
+ if (filePath) {
227
+ await announceSavedReport({ report, filePath, reportsDir: config.outputDir, values, lang });
228
+ }
229
+ const exitCode = await emitBatchVerdict(report, config.outputDir, values, lang);
230
+ throw new CliExit(applyGateExitCode(exitCode, values, lang));
231
+ }
232
+ let report;
233
+ let filePath;
234
+ if (repeatCount > 1 && !config.dryRun) {
235
+ const result = await runMultiple({
236
+ ...config,
237
+ repeat: repeatCount,
238
+ onRepeatProgress({ run, total }) {
239
+ process.stderr.write(tCli('cli.run.run_section', lang, { i: run, n: total }));
240
+ },
241
+ });
242
+ report = result.report;
243
+ filePath = null;
244
+ }
245
+ else {
246
+ const result = (await runEvaluation({
247
+ ...config,
248
+ ...(repeatCount > 1 ? { repeat: repeatCount } : {}),
249
+ }));
250
+ if (isDryRunReport(result.report)) {
251
+ console.log(JSON.stringify(result.report, null, 2));
252
+ console.log(tCli('cli.run.dry_run_no_scores', lang));
253
+ throw new CliExit(0);
254
+ }
255
+ report = result.report;
256
+ filePath = result.filePath;
257
+ }
258
+ // --gold-dir / eval.yaml goldDir: compute α/κ/Pearson against gold annotations and re-persist.
259
+ const goldDir = values['gold-dir'] ?? evalConfig?.goldDir;
260
+ if (goldDir && filePath) {
261
+ const { attachGoldAgreementToReport, formatGoldCompare } = await import('../../grading/gold-cli.js');
262
+ const out = attachGoldAgreementToReport({
263
+ report,
264
+ goldDir,
265
+ outputDir: config.outputDir,
266
+ samples: config.bootstrapSamples,
267
+ });
268
+ if (out.result && out.gold) {
269
+ process.stderr.write(formatGoldCompare(out.result, out.gold));
270
+ if (out.result.contaminationWarning) {
271
+ process.stderr.write(tCli('cli.run.contamination_warning', lang, {
272
+ warning: out.result.contaminationWarning,
273
+ }));
274
+ }
275
+ }
276
+ else {
277
+ process.stderr.write(tCli('cli.run.gold_load_failed', lang, { dir: goldDir }));
278
+ for (const m of out.loadIssues) {
279
+ process.stderr.write(tCli('cli.run.gold_load_issue', lang, { message: m }));
280
+ }
281
+ }
282
+ }
283
+ console.log(JSON.stringify(report, null, 2));
284
+ if (filePath) {
285
+ await announceSavedReport({ report, filePath, reportsDir: config.outputDir, values, lang });
286
+ }
287
+ const exitCode = await emitEvaluationVerdict(report, values);
288
+ throw new CliExit(applyGateExitCode(exitCode, values, lang));
289
+ }
290
+ catch (err) {
291
+ // CliExit 是显式 exit 信号(从 requireEvaluationReport 等子调用冒上来),
292
+ // 保持原 code 透传;只有真正运行时错误才包装成 CliExit(1)。
293
+ if (err instanceof CliExit)
294
+ throw err;
295
+ console.error(tCli('cli.common.error_prefix', lang, { message: err.message }));
296
+ throw new CliExit(1);
297
+ }
298
+ }
299
+ //# sourceMappingURL=eval-runner.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"eval-runner.js","sourceRoot":"","sources":["../../../../src/cli/commands/eval-runner.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,MAAM,gBAAgB,CAAC;AACzC,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,MAAM,YAAY,CAAC;AAChD,OAAO,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AACxD,OAAO,EAAE,cAAc,EAAE,MAAM,gBAAgB,CAAC;AAChD,OAAO,EAAE,eAAe,EAAE,MAAM,iBAAiB,CAAC;AAoBlD,SAAS,cAAc,CAAC,MAAe;IACrC,OAAO,OAAO,CAAC,MAAM,IAAI,OAAO,MAAM,KAAK,QAAQ,IAAK,MAA+B,CAAC,MAAM,KAAK,IAAI,CAAC,CAAC;AAC3G,CAAC;AAED,SAAS,mBAAmB,CAAC,MAAe;IAC1C,OAAO,cAAc,CAAC,MAAM,CAAC,IAAK,MAA8B,CAAC,KAAK,KAAK,IAAI,CAAC;AAClF,CAAC;AAED,SAAS,cAAc,CAAC,MAAoB;IAC1C,MAAM,YAAY,GAAG,MAAM,CAAC,SAA+B,CAAC;IAC5D,MAAM,aAAa,GAAG,YAAY,KAAK,SAAS,IAAI,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,YAAY,CAAC,CAAC;QACvF,CAAC,CAAC,MAAM,CAAC,YAAY,CAAC;QACtB,CAAC,CAAC,GAAG,CAAC;IACR,MAAM,UAAU,GAAG,MAAM,CAAC,cAAc,CAAuB,CAAC;IAChE,MAAM,kBAAkB,GAAG,UAAU,KAAK,SAAS,IAAI,MAAM,CAAC,QAAQ,CAAC,MAAM,CAAC,UAAU,CAAC,CAAC;QACxF,CAAC,CAAC,MAAM,CAAC,UAAU,CAAC;QACpB,CAAC,CAAC,SAAS,CAAC;IACd,OAAO,EAAE,aAAa,EAAE,kBAAkB,EAAE,CAAC;AAC/C,CAAC;AAED,SAAS,aAAa,CAAC,KAAa,EAAE,QAAgB;IACpD,OAAO,KAAK,KAAK,UAAU,IAAI,CAAC,KAAK,KAAK,MAAM,IAAI,QAAQ,CAAC,QAAQ,CAAC,MAAM,CAAC,CAAC,CAAC;AACjF,CAAC;AAED,SAAS,cAAc,CAAC,MAAoB;IAC1C,OAAO,MAAM,CAAC,aAAa,CAAC,KAAK,IAAI,IAAI,MAAM,CAAC,SAAS,CAAC,KAAK,IAAI,CAAC;AACtE,CAAC;AAED,SAAS,iBAAiB,CAAC,IAAY,EAAE,MAAoB,EAAE,IAAa;IAC1E,IAAI,CAAC,cAAc,CAAC,MAAM,CAAC;QAAE,OAAO,IAAI,CAAC;IACzC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,kCAAkC,EAAE,IAAI,CAAC,CAAC,CAAC;IACrE,OAAO,CAAC,CAAC;AACX,CAAC;AAED,KAAK,UAAU,qBAAqB,CAAC,MAAwB,EAAE,MAAoB;IACjF,MAAM,EAAE,cAAc,EAAE,iBAAiB,EAAE,GAAG,MAAM,MAAM,CAAC,4BAA4B,CAAC,CAAC;IACzF,MAAM,MAAM,GAAG,cAAc,CAAC,MAAM,EAAE,cAAc,CAAC,MAAM,CAAC,CAAC,CAAC;IAC9D,OAAO,CAAC,GAAG,CAAC,iBAAiB,CAAC,MAAM,EAAE,EAAE,OAAO,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC;IAC1D,OAAO,aAAa,CAAC,MAAM,CAAC,KAAK,EAAE,MAAM,CAAC,QAAQ,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;AAC9D,CAAC;AAED,SAAS,uBAAuB,CAC9B,KAA4B,EAC5B,IAA4C;IAE5C,OAAO;QACL,IAAI,EAAE,YAAY;QAClB,EAAE,EAAE,IAAI,CAAC,QAAQ;QACjB,IAAI,EAAE;YACJ,GAAG,KAAK,CAAC,IAAI;YACb,QAAQ,EAAE,CAAC,UAAU,EAAE,IAAI,CAAC,IAAI,CAAC;YACjC,WAAW,EAAE,IAAI,CAAC,WAAW;YAC7B,YAAY,EAAE,IAAI,CAAC,YAAY;YAC/B,cAAc,EAAE,IAAI,CAAC,YAAY,CAAC,CAAC,CAAC,EAAE,CAAC,IAAI,CAAC,IAAI,CAAC,EAAE,IAAI,CAAC,YAAY,EAAE,CAAC,CAAC,CAAC,EAAE;SAC5E;QACD,OAAO,EAAE,IAAI,CAAC,OAAO;QACrB,OAAO,EAAE,EAAE;QACX,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,IAAI,CAAC,QAAQ,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;KAClC,CAAC;AACxB,CAAC;AAED,KAAK,UAAU,qBAAqB,CAClC,KAA4B,EAC5B,UAAkB,EAClB,IAAa;IAEb,MAAM,EAAE,eAAe,EAAE,GAAG,MAAM,MAAM,CAAC,8BAA8B,CAAC,CAAC;IACzE,MAAM,KAAK,GAAG,eAAe,CAAC,UAAU,CAAC,CAAC;IAC1C,MAAM,OAAO,GAAuB,EAAE,CAAC;IACvC,KAAK,MAAM,IAAI,IAAI,KAAK,CAAC,KAAK,EAAE,CAAC;QAC/B,MAAM,MAAM,GAAG,MAAM,KAAK,CAAC,GAAG,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC;QAC9C,IAAI,MAAM,EAAE,IAAI,KAAK,YAAY,EAAE,CAAC;YAClC,OAAO,CAAC,IAAI,CAAC,MAAM,CAAC,CAAC;QACvB,CAAC;aAAM,CAAC;YACN,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,oCAAoC,EAAE,IAAI,EAAE,EAAE,EAAE,EAAE,IAAI,CAAC,QAAQ,EAAE,CAAC,CAAC,CAAC;YAC9F,OAAO,CAAC,IAAI,CAAC,uBAAuB,CAAC,KAAK,EAAE,IAAI,CAAC,CAAC,CAAC;QACrD,CAAC;IACH,CAAC;IACD,OAAO,OAAO,CAAC;AACjB,CAAC;AAED,KAAK,UAAU,gBAAgB,CAC7B,MAA6B,EAC7B,UAAkB,EAClB,MAAoB,EACpB,IAAa;IAEb,MAAM,EAAE,cAAc,EAAE,GAAG,MAAM,MAAM,CAAC,4BAA4B,CAAC,CAAC;IACtE,MAAM,YAAY,GAAG,MAAM,qBAAqB,CAAC,MAAM,EAAE,UAAU,EAAE,IAAI,CAAC,CAAC;IAC3E,MAAM,OAAO,GAAG,YAAY,CAAC,GAAG,CAAC,CAAC,KAAK,EAAE,EAAE,CAAC,CAAC;QAC3C,EAAE,EAAE,KAAK,CAAC,EAAE;QACZ,SAAS,EAAE,KAAK,CAAC,IAAI,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,KAAK,CAAC,EAAE;QAC7C,OAAO,EAAE,cAAc,CAAC,KAAK,EAAE,cAAc,CAAC,MAAM,CAAC,CAAC;KACvD,CAAC,CAAC,CAAC;IACJ,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,EAAE,CAAC,aAAa,CAAC,CAAC,CAAC,OAAO,CAAC,KAAK,EAAE,CAAC,CAAC,OAAO,CAAC,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC;IAChG,MAAM,MAAM,GAAG,OAAO,CAAC,MAAM,GAAG,MAAM,CAAC;IAEvC,MAAM,MAAM,GAAG,IAAI,KAAK,IAAI;QAC1B,CAAC,CAAC,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC,KAAK,CAAC;QAC/B,CAAC,CAAC,CAAC,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC;IACrC,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,8BAA8B,EAAE,IAAI,EAAE;QACrD,MAAM;QACN,MAAM;QACN,KAAK,EAAE,OAAO,CAAC,MAAM;KACtB,CAAC,CAAC,CAAC;IACJ,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,OAAO,CAAC,GAAG,CAAC,KAAK,MAAM,CAAC,OAAO,CAAC,KAAK,KAAK,MAAM,CAAC,SAAS,MAAM,MAAM,CAAC,OAAO,CAAC,QAAQ,EAAE,CAAC,CAAC;IAC7F,CAAC;IACD,OAAO,MAAM,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;AAC9B,CAAC;AAED,KAAK,UAAU,mBAAmB,CAAC,EACjC,MAAM,EACN,QAAQ,EACR,UAAU,EACV,MAAM,EACN,IAAI,GAOL;IACC,MAAM,KAAK,GAAG,eAAe,CAAC,MAAM,CAAC,CAAC;IACtC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,MAAM,CAAC,IAAI,KAAK,kBAAkB,CAAC,CAAC,CAAC,wBAAwB,CAAC,CAAC,CAAC,uBAAuB,EAAE,IAAI,CAAC,CAAC,CAAC;IAC1H,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,eAAe,EAAE,IAAI,EAAE,KAAK,CAAC,CAAC,CAAC;IACzD,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,sBAAsB,EAAE,IAAI,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,CAAC,CAAC,CAAC;IAE7E,IAAI,CAAC,MAAM,CAAC,UAAU,CAAC,IAAI,OAAO,CAAC,MAAM,CAAC,KAAK,EAAE,CAAC;QAChD,MAAM,EAAE,kBAAkB,EAAE,GAAG,MAAM,MAAM,CAAC,+BAA+B,CAAC,CAAC;QAC7E,MAAM,MAAM,GAAiB,kBAAkB,CAAC,EAAE,UAAU,EAAE,CAAC,CAAC;QAChE,MAAM,SAAS,GAAW,MAAM,MAAM,CAAC,KAAK,EAAE,CAAC;QAC/C,MAAM,SAAS,GAAW,GAAG,SAAS,YAAY,MAAM,CAAC,EAAE,EAAE,CAAC;QAC9D,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,+BAA+B,EAAE,IAAI,EAAE,EAAE,GAAG,EAAE,SAAS,EAAE,CAAC,CAAC,CAAC;QACtF,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,4BAA4B,EAAE,IAAI,EAAE,EAAE,GAAG,EAAE,SAAS,EAAE,CAAC,CAAC,CAAC;QACnF,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,4BAA4B,EAAE,IAAI,CAAC,CAAC,CAAC;QAE/D,MAAM,EAAE,QAAQ,EAAE,GAAG,MAAM,MAAM,CAAC,SAAS,CAAC,CAAC;QAC7C,MAAM,OAAO,GAAW,QAAQ,EAAE,KAAK,QAAQ,CAAC,CAAC,CAAC,MAAM,CAAC,CAAC,CAAC,QAAQ,EAAE,KAAK,OAAO,CAAC,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,UAAU,CAAC;QACzG,MAAM,EAAE,QAAQ,EAAE,UAAU,EAAE,GAAG,MAAM,MAAM,CAAC,oBAAoB,CAAC,CAAC;QACpE,UAAU,CAAC,OAAO,EAAE,CAAC,SAAS,CAAC,EAAE,GAAG,EAAE,CAAC,SAAS,CAAC,CAAC;IACpD,CAAC;SAAM,IAAI,CAAC,MAAM,CAAC,UAAU,CAAC,EAAE,CAAC;QAC/B,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,6BAA6B,EAAE,IAAI,CAAC,CAAC,CAAC;QAChE,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,4BAA4B,EAAE,IAAI,EAAE,EAAE,EAAE,EAAE,MAAM,CAAC,EAAE,EAAE,GAAG,EAAE,UAAU,EAAE,CAAC,CAAC,CAAC;IACrG,CAAC;AACH,CAAC;AAED,MAAM,CAAC,KAAK,UAAU,OAAO,CAAC,IAAc;IAC1C,MAAM,IAAI,GAAG,YAAY,CAAC,IAAI,CAAC,CAAC;IAChC,iEAAiE;IACjE,uFAAuF;IACvF,0CAA0C;IAC1C,MAAM,EAAE,MAAM,EAAE,MAAM,EAAE,UAAU,EAAE,GAAG,cAAc,CAAC,IAAI,EAAE;QAC1D,KAAK,EAAE,EAAE,IAAI,EAAE,SAAS,EAAE;QAC1B,MAAM,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE;QAC1B,cAAc,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE;QAClC,SAAS,EAAE,EAAE,IAAI,EAAE,SAAS,EAAE;QAC9B,mBAAmB,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE;QACvC,UAAU,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE;QAC9B,kBAAkB,EAAE,EAAE,IAAI,EAAE,SAAS,EAAE;QACvC,YAAY,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE;QAChC,uBAAuB,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE;QAC3C,sBAAsB,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE;QAC1C,SAAS,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,KAAK,EAAE;QAC7C,cAAc,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE;QAClC,aAAa,EAAE,EAAE,IAAI,EAAE,SAAS,EAAE;QAClC,SAAS,EAAE,EAAE,IAAI,EAAE,SAAS,EAAE;KAC/B,CAAC,CAAC;IAEH,MAAM,EAAE,aAAa,EAAE,WAAW,EAAE,kBAAkB,EAAE,GAAG,MAAM,MAAM,CAAC,wCAAwC,CAAC,CAAC;IAElH,IAAI,MAAM,CAAC,KAAK,KAAK,SAAS,EAAE,CAAC;QAC/B,MAAM,CAAC,KAAK,GAAG,MAAM,CAAC,KAA4B,CAAC;IACrD,CAAC;IACD,MAAM,CAAC,UAAU,GAAG,cAAc,CAAC,IAAI,CAAgC,CAAC;IAExE,kDAAkD;IAClD,MAAM,SAAS,GAAG,MAAM,CAAC,MAA4B,CAAC;IACtD,MAAM,YAAY,GAAG,SAAS,KAAK,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,SAAS,CAAC,CAAC,CAAC,CAAC,CAAC,UAAU,EAAE,MAAM,IAAI,CAAC,CAAC,CAAC;IAC7F,IAAI,SAAS,KAAK,SAAS,IAAI,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,YAAY,CAAC,IAAI,YAAY,GAAG,CAAC,CAAC,EAAE,CAAC;QACpF,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,wBAAwB,EAAE,IAAI,EAAE,EAAE,KAAK,EAAE,SAAS,EAAE,CAAC,CAAC,CAAC;IACnF,CAAC;IACD,MAAM,WAAW,GAAW,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,YAAY,CAAC,IAAI,CAAC,CAAC,CAAC;IAEvE,uCAAuC;IACvC,MAAM,cAAc,GAAG,MAAM,CAAC,cAAc,CAAuB,CAAC;IACpE,MAAM,iBAAiB,GAAG,cAAc,KAAK,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,cAAc,CAAC,CAAC,CAAC,CAAC,CAAC,UAAU,EAAE,WAAW,IAAI,CAAC,CAAC,CAAC;IACjH,IAAI,cAAc,KAAK,SAAS,IAAI,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,iBAAiB,CAAC,IAAI,iBAAiB,GAAG,CAAC,CAAC,EAAE,CAAC;QACnG,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,8BAA8B,EAAE,IAAI,EAAE,EAAE,KAAK,EAAE,cAAc,EAAE,CAAC,CAAC,CAAC;IAC9F,CAAC;IACD,MAAM,gBAAgB,GAAW,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,IAAI,CAAC,KAAK,CAAC,iBAAiB,CAAC,IAAI,CAAC,CAAC,CAAC;IACjF,IAAI,gBAAgB,GAAG,CAAC;QAAE,MAAM,CAAC,WAAW,GAAG,gBAAgB,CAAC;IAEhE,mEAAmE;IACnE,qEAAqE;IACrE,uEAAuE;IACvE,8DAA8D;IAC9D,MAAM,SAAS,GAAG,MAAM,CAAC,YAAY,CAAC,IAAI,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,YAAY,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;IAC1F,MAAM,kBAAkB,GAAG,MAAM,CAAC,uBAAuB,CAAC,IAAI,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,uBAAuB,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;IACzH,MAAM,iBAAiB,GAAG,MAAM,CAAC,sBAAsB,CAAC,IAAI,IAAI,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,sBAAsB,CAAC,CAAC,CAAC,CAAC,CAAC,SAAS,CAAC;IACtH,IAAI,SAAS,KAAK,SAAS,IAAI,kBAAkB,KAAK,SAAS,IAAI,iBAAiB,KAAK,SAAS,EAAE,CAAC;QACnG,MAAM,CAAC,MAAM,GAAG;YACd,GAAG,CAAC,SAAS,KAAK,SAAS,IAAI,MAAM,CAAC,QAAQ,CAAC,SAAS,CAAC,IAAI,SAAS,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,QAAQ,EAAE,SAAS,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;YAC3G,GAAG,CAAC,kBAAkB,KAAK,SAAS,IAAI,MAAM,CAAC,QAAQ,CAAC,kBAAkB,CAAC,IAAI,kBAAkB,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,YAAY,EAAE,kBAAkB,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;YACnJ,GAAG,CAAC,iBAAiB,KAAK,SAAS,IAAI,MAAM,CAAC,QAAQ,CAAC,iBAAiB,CAAC,IAAI,iBAAiB,IAAI,CAAC,CAAC,CAAC,CAAC,EAAE,WAAW,EAAE,iBAAiB,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;SAC/I,CAAC;IACJ,CAAC;IAED,6FAA6F;IAC7F,iGAAiG;IACjG,gEAAgE;IAChE,MAAM,eAAe,GAAI,MAAM,CAAC,kBAAkB,CAAyB,KAAK,IAAI;WAC/E,CAAC,MAAM,CAAC,kBAAkB,CAAC,KAAK,SAAS,IAAI,UAAU,EAAE,YAAY,KAAK,KAAK,CAAC,CAAC;IACtF,IAAI,eAAe,EAAE,CAAC;QACpB,MAAM,CAAC,YAAY,GAAG,KAAK,CAAC;QAC5B,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,iCAAiC,EAAE,IAAI,CAAC,CAAC,CAAC;IACtE,CAAC;IAED,2EAA2E;IAC3E,6EAA6E;IAC7E,8DAA8D;IAC9D,MAAM,gBAAgB,GAAI,MAAM,CAAC,SAAiC,KAAK,IAAI;WACtE,CAAC,MAAM,CAAC,SAAS,KAAK,SAAS,IAAI,UAAU,EAAE,SAAS,KAAK,KAAK,CAAC,CAAC;IACzE,IAAI,gBAAgB,EAAE,CAAC;QACrB,MAAM,CAAC,SAAS,GAAG,IAAI,CAAC;QACxB,MAAM,KAAK,GAAG,MAAM,CAAC,mBAAmB,CAAuB,CAAC;QAChE,MAAM,QAAQ,GAAG,KAAK,KAAK,SAAS,CAAC,CAAC,CAAC,MAAM,CAAC,KAAK,CAAC,CAAC,CAAC,CAAC,CAAC,UAAU,EAAE,gBAAgB,IAAI,IAAI,CAAC,CAAC;QAC9F,IAAI,KAAK,KAAK,SAAS,IAAI,CAAC,CAAC,MAAM,CAAC,QAAQ,CAAC,QAAQ,CAAC,IAAI,QAAQ,GAAG,GAAG,CAAC,EAAE,CAAC;YAC1E,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,mCAAmC,EAAE,IAAI,EAAE,EAAE,KAAK,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC;QAC1F,CAAC;QACD,MAAM,OAAO,GAAG,IAAI,CAAC,GAAG,CAAC,GAAG,EAAE,IAAI,CAAC,KAAK,CAAC,QAAQ,CAAC,IAAI,IAAI,CAAC,CAAC;QAC5D,IAAI,OAAO,GAAG,KAAK,EAAE,CAAC;YACpB,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,qCAAqC,EAAE,IAAI,EAAE,EAAE,CAAC,EAAE,OAAO,EAAE,CAAC,CAAC,CAAC;QAC1F,CAAC;QACD,MAAM,CAAC,gBAAgB,GAAG,OAAO,CAAC;IACpC,CAAC;IAED,oDAAoD;IACpD,MAAM,CAAC,IAAI,GAAG,IAAI,CAAC;IACnB,IAAI,MAAM,CAAC,mBAAmB,CAAY,EAAE,CAAC;QAC3C,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,mCAAmC,EAAE,IAAI,CAAC,GAAG,IAAI,CAAC,CAAC;IAC/E,CAAC;IAED,IAAI,CAAC;QACH,kDAAkD;QAClD,IAAI,MAAM,CAAC,KAAK,EAAE,CAAC;YACjB,MAAM,EAAE,MAAM,EAAE,QAAQ,EAAE,GAAG,MAAM,kBAAkB,CAAC;gBACpD,GAAG,MAAM;gBACT,MAAM,EAAE,WAAW;gBACnB,eAAe,CAAC,EAAE,KAAK,EAAE,KAAK,EAAE,OAAO,EAAE,KAAK,EAAqB;oBACjE,IAAI,KAAK,KAAK,OAAO,EAAE,CAAC;wBACtB,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,uBAAuB,EAAE,IAAI,EAAE;4BACvD,CAAC,EAAE,OAAO,IAAI,EAAE,EAAE,CAAC,EAAE,KAAK,IAAI,EAAE,EAAE,KAAK,EAAE,KAAK,IAAI,EAAE;yBACrD,CAAC,CAAC,CAAC;oBACN,CAAC;gBACH,CAAC;aACF,CAAmF,CAAC;YACrF,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC;YAC7C,IAAI,mBAAmB,CAAC,MAAM,CAAC,EAAE,CAAC;gBAChC,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,2BAA2B,EAAE,IAAI,CAAC,CAAC,CAAC;gBACrD,MAAM,IAAI,OAAO,CAAC,CAAC,CAAC,CAAC;YACvB,CAAC;YACD,IAAI,QAAQ,EAAE,CAAC;gBACb,MAAM,mBAAmB,CAAC,EAAE,MAAM,EAAE,QAAQ,EAAE,UAAU,EAAE,MAAM,CAAC,SAAS,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC;YAC9F,CAAC;YACD,MAAM,QAAQ,GAAG,MAAM,gBAAgB,CAAC,MAAM,EAAE,MAAM,CAAC,SAAS,EAAE,MAAM,EAAE,IAAI,CAAC,CAAC;YAChF,MAAM,IAAI,OAAO,CAAC,iBAAiB,CAAC,QAAQ,EAAE,MAAM,EAAE,IAAI,CAAC,CAAC,CAAC;QAC/D,CAAC;QAED,IAAI,MAAc,CAAC;QACnB,IAAI,QAAuB,CAAC;QAE5B,IAAI,WAAW,GAAG,CAAC,IAAI,CAAC,MAAM,CAAC,MAAM,EAAE,CAAC;YACtC,MAAM,MAAM,GAAG,MAAM,WAAW,CAAC;gBAC/B,GAAG,MAAM;gBACT,MAAM,EAAE,WAAW;gBACnB,gBAAgB,CAAC,EAAE,GAAG,EAAE,KAAK,EAAsB;oBACjD,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,qBAAqB,EAAE,IAAI,EAAE,EAAE,CAAC,EAAE,GAAG,EAAE,CAAC,EAAE,KAAK,EAAE,CAAC,CAAC,CAAC;gBAChF,CAAC;aACF,CAAuB,CAAC;YACzB,MAAM,GAAG,MAAM,CAAC,MAAM,CAAC;YACvB,QAAQ,GAAG,IAAI,CAAC;QAClB,CAAC;aAAM,CAAC;YACN,MAAM,MAAM,GAAG,CAAC,MAAM,aAAa,CAAC;gBAClC,GAAG,MAAM;gBACT,GAAG,CAAC,WAAW,GAAG,CAAC,CAAC,CAAC,CAAC,EAAE,MAAM,EAAE,WAAW,EAAE,CAAC,CAAC,CAAC,EAAE,CAAC;aACpD,CAAC,CAAe,CAAC;YAClB,IAAI,cAAc,CAAC,MAAM,CAAC,MAAM,CAAC,EAAE,CAAC;gBAClC,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC;gBACpD,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,2BAA2B,EAAE,IAAI,CAAC,CAAC,CAAC;gBACrD,MAAM,IAAI,OAAO,CAAC,CAAC,CAAC,CAAC;YACvB,CAAC;YACD,MAAM,GAAG,MAAM,CAAC,MAAgB,CAAC;YACjC,QAAQ,GAAG,MAAM,CAAC,QAAQ,CAAC;QAC7B,CAAC;QAED,+FAA+F;QAC/F,MAAM,OAAO,GAAI,MAAM,CAAC,UAAU,CAAwB,IAAI,UAAU,EAAE,OAAO,CAAC;QAClF,IAAI,OAAO,IAAI,QAAQ,EAAE,CAAC;YACxB,MAAM,EAAE,2BAA2B,EAAE,iBAAiB,EAAE,GAAG,MAAM,MAAM,CAAC,2BAA2B,CAAC,CAAC;YACrG,MAAM,GAAG,GAAG,2BAA2B,CAAC;gBACtC,MAAM;gBACN,OAAO;gBACP,SAAS,EAAE,MAAM,CAAC,SAAS;gBAC3B,OAAO,EAAE,MAAM,CAAC,gBAAgB;aACjC,CAAC,CAAC;YACH,IAAI,GAAG,CAAC,MAAM,IAAI,GAAG,CAAC,IAAI,EAAE,CAAC;gBAC3B,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,iBAAiB,CAAC,GAAG,CAAC,MAAM,EAAE,GAAG,CAAC,IAAI,CAAC,CAAC,CAAC;gBAC9D,IAAI,GAAG,CAAC,MAAM,CAAC,oBAAoB,EAAE,CAAC;oBACpC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,+BAA+B,EAAE,IAAI,EAAE;wBAC/D,OAAO,EAAE,GAAG,CAAC,MAAM,CAAC,oBAAoB;qBACzC,CAAC,CAAC,CAAC;gBACN,CAAC;YACH,CAAC;iBAAM,CAAC;gBACN,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,0BAA0B,EAAE,IAAI,EAAE,EAAE,GAAG,EAAE,OAAO,EAAE,CAAC,CAAC,CAAC;gBAC/E,KAAK,MAAM,CAAC,IAAI,GAAG,CAAC,UAAU,EAAE,CAAC;oBAC/B,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,yBAAyB,EAAE,IAAI,EAAE,EAAE,OAAO,EAAE,CAAC,EAAE,CAAC,CAAC,CAAC;gBAC9E,CAAC;YACH,CAAC;QACH,CAAC;QAED,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC;QAC7C,IAAI,QAAQ,EAAE,CAAC;YACb,MAAM,mBAAmB,CAAC,EAAE,MAAM,EAAE,QAAQ,EAAE,UAAU,EAAE,MAAM,CAAC,SAAS,EAAE,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC;QAC9F,CAAC;QACD,MAAM,QAAQ,GAAG,MAAM,qBAAqB,CAAC,MAAM,EAAE,MAAM,CAAC,CAAC;QAC7D,MAAM,IAAI,OAAO,CAAC,iBAAiB,CAAC,QAAQ,EAAE,MAAM,EAAE,IAAI,CAAC,CAAC,CAAC;IAC/D,CAAC;IAAC,OAAO,GAAY,EAAE,CAAC;QACtB,0DAA0D;QAC1D,wCAAwC;QACxC,IAAI,GAAG,YAAY,OAAO;YAAE,MAAM,GAAG,CAAC;QACtC,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,yBAAyB,EAAE,IAAI,EAAE,EAAE,OAAO,EAAG,GAAa,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC;QAC1F,MAAM,IAAI,OAAO,CAAC,CAAC,CAAC,CAAC;IACvB,CAAC;AACH,CAAC"}
@@ -0,0 +1,2 @@
1
+ export declare function execute(argv: string[]): Promise<void>;
2
+ //# sourceMappingURL=eval.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"eval.d.ts","sourceRoot":"","sources":["../../../../src/cli/commands/eval.ts"],"names":[],"mappings":"AAGA,wBAAsB,OAAO,CAAC,IAAI,EAAE,MAAM,EAAE,GAAG,OAAO,CAAC,IAAI,CAAC,CAQ3D"}
@@ -0,0 +1,11 @@
1
+ import { execute as gold } from './eval-gold.js';
2
+ import { execute as run } from './eval-runner.js';
3
+ export async function execute(argv) {
4
+ const [sub, ...rest] = argv;
5
+ if (sub === 'gold') {
6
+ await gold(rest);
7
+ return;
8
+ }
9
+ await run(argv);
10
+ }
11
+ //# sourceMappingURL=eval.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"eval.js","sourceRoot":"","sources":["../../../../src/cli/commands/eval.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,IAAI,IAAI,EAAE,MAAM,gBAAgB,CAAC;AACjD,OAAO,EAAE,OAAO,IAAI,GAAG,EAAE,MAAM,kBAAkB,CAAC;AAElD,MAAM,CAAC,KAAK,UAAU,OAAO,CAAC,IAAc;IAC1C,MAAM,CAAC,GAAG,EAAE,GAAG,IAAI,CAAC,GAAG,IAAI,CAAC;IAC5B,IAAI,GAAG,KAAK,MAAM,EAAE,CAAC;QACnB,MAAM,IAAI,CAAC,IAAI,CAAC,CAAC;QACjB,OAAO;IACT,CAAC;IAED,MAAM,GAAG,CAAC,IAAI,CAAC,CAAC;AAClB,CAAC"}
@@ -0,0 +1,2 @@
1
+ export declare function execute(argv: string[]): Promise<void>;
2
+ //# sourceMappingURL=evolve.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"evolve.d.ts","sourceRoot":"","sources":["../../../../src/cli/commands/evolve.ts"],"names":[],"mappings":"AA0CA,wBAAsB,OAAO,CAAC,IAAI,EAAE,MAAM,EAAE,GAAG,OAAO,CAAC,IAAI,CAAC,CA4G3D"}
@@ -0,0 +1,115 @@
1
+ import { CliExit } from '../cli-exit.js';
2
+ import { resolve, join } from 'node:path';
3
+ import { existsSync } from 'node:fs';
4
+ import { tCli, langFromArgv } from '../i18n.js';
5
+ import { COMMON_OPTIONS } from '../parse-run-config.js';
6
+ import { parseArgsStrictOrExit } from '../parse-strict.js';
7
+ import { makeOnProgress } from '../progress.js';
8
+ export async function execute(argv) {
9
+ const lang = langFromArgv(argv);
10
+ const { values, positionals } = parseArgsStrictOrExit({
11
+ args: argv,
12
+ options: {
13
+ ...COMMON_OPTIONS,
14
+ rounds: { type: 'string', default: '5' },
15
+ target: { type: 'string' },
16
+ samples: { type: 'string', default: 'eval-samples.json' },
17
+ model: { type: 'string', default: 'sonnet' },
18
+ 'judge-models': { type: 'string', default: 'claude:haiku' },
19
+ 'improve-model': { type: 'string', default: 'sonnet' },
20
+ concurrency: { type: 'string', default: '1' },
21
+ timeout: { type: 'string', default: '120' },
22
+ executor: { type: 'string', default: 'claude' },
23
+ 'skip-connectivity': { type: 'boolean', default: false },
24
+ },
25
+ allowPositionals: true,
26
+ });
27
+ // skill path 走 parseArgs 的 positionals,避免 raw argv.find 把 flag value
28
+ // 当成 path 误识别,例如 `omk evolve --judge-models openai-api:gpt-4o foo.md`。
29
+ const skillPath = positionals[0];
30
+ if (!skillPath) {
31
+ console.error(tCli('cli.evolve.specify_skill_path', lang));
32
+ throw new CliExit(1);
33
+ }
34
+ let samplesFile = values.samples ?? 'eval-samples.json';
35
+ if (samplesFile === 'eval-samples.json' && !existsSync(resolve(samplesFile))) {
36
+ if (existsSync(resolve('eval-samples.yaml')))
37
+ samplesFile = 'eval-samples.yaml';
38
+ else if (existsSync(resolve('eval-samples.yml')))
39
+ samplesFile = 'eval-samples.yml';
40
+ }
41
+ const { evolveSkill } = await import('../../authoring/evolver.js');
42
+ const { parseJudgeModelsArgOrExit } = await import('../parse-run-config.js');
43
+ const evolveJudges = parseJudgeModelsArgOrExit(values['judge-models']);
44
+ if (evolveJudges.length > 1) {
45
+ console.error(tCli('cli.common.judge_models_single_only', lang, { cmd: 'omk evolve' }));
46
+ throw new CliExit(2);
47
+ }
48
+ process.stderr.write(tCli('cli.evolve.section_header', lang, { path: skillPath }));
49
+ try {
50
+ const result = await evolveSkill({
51
+ skillPath: resolve(skillPath),
52
+ samplesPath: resolve(samplesFile),
53
+ rounds: Math.max(1, Number(values.rounds) || 5),
54
+ target: values.target ? Number(values.target) : null,
55
+ model: values.model,
56
+ judgeModels: evolveJudges,
57
+ improveModel: values['improve-model'],
58
+ executorName: values.executor,
59
+ concurrency: Math.max(1, Number(values.concurrency) || 1),
60
+ timeoutMs: Math.max(1, Number(values.timeout) || 120) * 1000,
61
+ skipConnectivity: values['skip-connectivity'],
62
+ onProgress: makeOnProgress(lang),
63
+ onRoundProgress({ round, totalRounds: _totalRounds, phase, score, delta, accepted, costUSD, costReported, error }) {
64
+ // costReported=false 时显示「—」而不是 $0.0000(executor 不报 cost,如 codex)。
65
+ // 缺位 / true 当 reported 走旧格式。
66
+ const fmtRoundCost = (c, r) => r ? `$${c.toFixed(4)}` : '—';
67
+ if (phase === 'baseline') {
68
+ process.stderr.write(tCli('cli.evolve.round_baseline', lang, {
69
+ score: score.toFixed(2), cost: fmtRoundCost(costUSD, costReported !== false),
70
+ }));
71
+ }
72
+ else if (phase === 'error') {
73
+ process.stderr.write(tCli('cli.evolve.round_error', lang, {
74
+ round, error: String(error ?? ''),
75
+ }));
76
+ }
77
+ else if (phase === 'done') {
78
+ const delta_ = delta >= 0 ? `+${delta.toFixed(2)}` : delta.toFixed(2);
79
+ const status = accepted ? '✓ ACCEPT' : '✗ REJECT';
80
+ process.stderr.write(tCli('cli.evolve.round_done', lang, {
81
+ round, score: score.toFixed(2), delta: delta_, status, cost: fmtRoundCost(costUSD, costReported !== false),
82
+ }));
83
+ }
84
+ },
85
+ });
86
+ const improvement = result.startScore > 0
87
+ ? ((result.finalScore - result.startScore) / result.startScore * 100).toFixed(1)
88
+ : '0';
89
+ const totalCostStr = result.costReported === false
90
+ ? '—' // 任一轮的 executor 不报 cost → totalCostUSD 是 lower-bound
91
+ : `$${result.totalCostUSD.toFixed(4)}`;
92
+ process.stderr.write(tCli('cli.evolve.summary', lang, {
93
+ start: result.startScore.toFixed(2), final: result.finalScore.toFixed(2),
94
+ percent: improvement, rounds: result.totalRounds, cost: totalCostStr,
95
+ }));
96
+ process.stderr.write(tCli('cli.evolve.best_path', lang, {
97
+ best: result.bestSkillPath, target: resolve(skillPath),
98
+ }));
99
+ process.stderr.write(tCli('cli.evolve.versions_saved', lang, {
100
+ dir: join(resolve(skillPath, '..'), 'evolve'),
101
+ }));
102
+ if (result.reportId) {
103
+ process.stderr.write(tCli('cli.evolve.report_link', lang, { id: result.reportId }));
104
+ }
105
+ console.log(JSON.stringify(result, null, 2));
106
+ }
107
+ catch (err) {
108
+ // CliExit 是显式 exit 信号,保持原 code 透传(同 gate / run)。
109
+ if (err instanceof CliExit)
110
+ throw err;
111
+ console.error(tCli('cli.common.error_prefix', lang, { message: err.message }));
112
+ throw new CliExit(1);
113
+ }
114
+ }
115
+ //# sourceMappingURL=evolve.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"evolve.js","sourceRoot":"","sources":["../../../../src/cli/commands/evolve.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,MAAM,gBAAgB,CAAC;AACzC,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAC1C,OAAO,EAAE,UAAU,EAAE,MAAM,SAAS,CAAC;AACrC,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,MAAM,YAAY,CAAC;AAChD,OAAO,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AACxD,OAAO,EAAE,qBAAqB,EAAE,MAAM,oBAAoB,CAAC;AAC3D,OAAO,EAAE,cAAc,EAAE,MAAM,gBAAgB,CAAC;AAoChD,MAAM,CAAC,KAAK,UAAU,OAAO,CAAC,IAAc;IAC1C,MAAM,IAAI,GAAG,YAAY,CAAC,IAAI,CAAC,CAAC;IAChC,MAAM,EAAE,MAAM,EAAE,WAAW,EAAE,GAAG,qBAAqB,CAAC;QACpD,IAAI,EAAE,IAAI;QACV,OAAO,EAAE;YACP,GAAG,cAAc;YACjB,MAAM,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,GAAG,EAAE;YACxC,MAAM,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE;YAC1B,OAAO,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,mBAAmB,EAAE;YACzD,KAAK,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,QAAQ,EAAE;YAC5C,cAAc,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,cAAc,EAAE;YAC3D,eAAe,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,QAAQ,EAAE;YACtD,WAAW,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,GAAG,EAAE;YAC7C,OAAO,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,KAAK,EAAE;YAC3C,QAAQ,EAAE,EAAE,IAAI,EAAE,QAAQ,EAAE,OAAO,EAAE,QAAQ,EAAE;YAC/C,mBAAmB,EAAE,EAAE,IAAI,EAAE,SAAS,EAAE,OAAO,EAAE,KAAK,EAAE;SACzD;QACD,gBAAgB,EAAE,IAAI;KACvB,CAAC,CAAC;IAEH,qEAAqE;IACrE,uEAAuE;IACvE,MAAM,SAAS,GAAuB,WAAW,CAAC,CAAC,CAAC,CAAC;IACrD,IAAI,CAAC,SAAS,EAAE,CAAC;QACf,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,+BAA+B,EAAE,IAAI,CAAC,CAAC,CAAC;QAC3D,MAAM,IAAI,OAAO,CAAC,CAAC,CAAC,CAAC;IACvB,CAAC;IAED,IAAI,WAAW,GAAY,MAAM,CAAC,OAAkB,IAAI,mBAAmB,CAAC;IAC5E,IAAI,WAAW,KAAK,mBAAmB,IAAI,CAAC,UAAU,CAAC,OAAO,CAAC,WAAW,CAAC,CAAC,EAAE,CAAC;QAC7E,IAAI,UAAU,CAAC,OAAO,CAAC,mBAAmB,CAAC,CAAC;YAAE,WAAW,GAAG,mBAAmB,CAAC;aAC3E,IAAI,UAAU,CAAC,OAAO,CAAC,kBAAkB,CAAC,CAAC;YAAE,WAAW,GAAG,kBAAkB,CAAC;IACrF,CAAC;IAED,MAAM,EAAE,WAAW,EAAE,GAAG,MAAM,MAAM,CAAC,4BAA4B,CAAC,CAAC;IACnE,MAAM,EAAE,yBAAyB,EAAE,GAAG,MAAM,MAAM,CAAC,wBAAwB,CAAC,CAAC;IAE7E,MAAM,YAAY,GAAG,yBAAyB,CAAC,MAAM,CAAC,cAAc,CAAW,CAAC,CAAC;IACjF,IAAI,YAAY,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC5B,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,qCAAqC,EAAE,IAAI,EAAE,EAAE,GAAG,EAAE,YAAY,EAAE,CAAC,CAAC,CAAC;QACxF,MAAM,IAAI,OAAO,CAAC,CAAC,CAAC,CAAC;IACvB,CAAC;IAED,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,2BAA2B,EAAE,IAAI,EAAE,EAAE,IAAI,EAAE,SAAS,EAAE,CAAC,CAAC,CAAC;IAEnF,IAAI,CAAC;QACH,MAAM,MAAM,GAAiB,MAAM,WAAW,CAAC;YAC7C,SAAS,EAAE,OAAO,CAAC,SAAS,CAAC;YAC7B,WAAW,EAAE,OAAO,CAAC,WAAW,CAAC;YACjC,MAAM,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;YAC/C,MAAM,EAAE,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,MAAM,CAAC,MAAM,CAAC,MAAM,CAAC,CAAC,CAAC,CAAC,IAAI;YACpD,KAAK,EAAE,MAAM,CAAC,KAAe;YAC7B,WAAW,EAAE,YAAY;YACzB,YAAY,EAAE,MAAM,CAAC,eAAe,CAAW;YAC/C,YAAY,EAAE,MAAM,CAAC,QAAkB;YACvC,WAAW,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,MAAM,CAAC,MAAM,CAAC,WAAW,CAAC,IAAI,CAAC,CAAC;YACzD,SAAS,EAAE,IAAI,CAAC,GAAG,CAAC,CAAC,EAAE,MAAM,CAAC,MAAM,CAAC,OAAO,CAAC,IAAI,GAAG,CAAC,GAAG,IAAI;YAC5D,gBAAgB,EAAE,MAAM,CAAC,mBAAmB,CAAY;YACxD,UAAU,EAAE,cAAc,CAAC,IAAI,CAAgC;YAC/D,eAAe,CAAC,EAAE,KAAK,EAAE,WAAW,EAAE,YAAY,EAAE,KAAK,EAAE,KAAK,EAAE,KAAK,EAAE,QAAQ,EAAE,OAAO,EAAE,YAAY,EAAE,KAAK,EAAqB;gBAClI,kEAAkE;gBAClE,6BAA6B;gBAC7B,MAAM,YAAY,GAAG,CAAC,CAAS,EAAE,CAAU,EAAU,EAAE,CAAC,CAAC,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,GAAG,CAAC;gBACrF,IAAI,KAAK,KAAK,UAAU,EAAE,CAAC;oBACzB,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,2BAA2B,EAAE,IAAI,EAAE;wBAC3D,KAAK,EAAE,KAAM,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,IAAI,EAAE,YAAY,CAAC,OAAQ,EAAE,YAAY,KAAK,KAAK,CAAC;qBAC/E,CAAC,CAAC,CAAC;gBACN,CAAC;qBAAM,IAAI,KAAK,KAAK,OAAO,EAAE,CAAC;oBAC7B,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,wBAAwB,EAAE,IAAI,EAAE;wBACxD,KAAK,EAAE,KAAK,EAAE,MAAM,CAAC,KAAK,IAAI,EAAE,CAAC;qBAClC,CAAC,CAAC,CAAC;gBACN,CAAC;qBAAM,IAAI,KAAK,KAAK,MAAM,EAAE,CAAC;oBAC5B,MAAM,MAAM,GAAW,KAAM,IAAI,CAAC,CAAC,CAAC,CAAC,IAAI,KAAM,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC,CAAC,KAAM,CAAC,OAAO,CAAC,CAAC,CAAC,CAAC;oBACjF,MAAM,MAAM,GAAW,QAAQ,CAAC,CAAC,CAAC,UAAU,CAAC,CAAC,CAAC,UAAU,CAAC;oBAC1D,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,uBAAuB,EAAE,IAAI,EAAE;wBACvD,KAAK,EAAE,KAAK,EAAE,KAAM,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,MAAM,EAAE,MAAM,EAAE,IAAI,EAAE,YAAY,CAAC,OAAQ,EAAE,YAAY,KAAK,KAAK,CAAC;qBAC7G,CAAC,CAAC,CAAC;gBACN,CAAC;YACH,CAAC;SACF,CAAC,CAAC;QAEH,MAAM,WAAW,GAAW,MAAM,CAAC,UAAU,GAAG,CAAC;YAC/C,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,UAAU,GAAG,MAAM,CAAC,UAAU,CAAC,GAAG,MAAM,CAAC,UAAU,GAAG,GAAG,CAAC,CAAC,OAAO,CAAC,CAAC,CAAC;YAChF,CAAC,CAAC,GAAG,CAAC;QACR,MAAM,YAAY,GAAG,MAAM,CAAC,YAAY,KAAK,KAAK;YAChD,CAAC,CAAC,GAAG,CAAE,qDAAqD;YAC5D,CAAC,CAAC,IAAI,MAAM,CAAC,YAAY,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,CAAC;QACzC,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,oBAAoB,EAAE,IAAI,EAAE;YACpD,KAAK,EAAE,MAAM,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC,CAAC,EAAE,KAAK,EAAE,MAAM,CAAC,UAAU,CAAC,OAAO,CAAC,CAAC,CAAC;YACxE,OAAO,EAAE,WAAW,EAAE,MAAM,EAAE,MAAM,CAAC,WAAW,EAAE,IAAI,EAAE,YAAY;SACrE,CAAC,CAAC,CAAC;QACJ,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,sBAAsB,EAAE,IAAI,EAAE;YACtD,IAAI,EAAE,MAAM,CAAC,aAAa,EAAE,MAAM,EAAE,OAAO,CAAC,SAAS,CAAC;SACvD,CAAC,CAAC,CAAC;QACJ,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,2BAA2B,EAAE,IAAI,EAAE;YAC3D,GAAG,EAAE,IAAI,CAAC,OAAO,CAAC,SAAS,EAAE,IAAI,CAAC,EAAE,QAAQ,CAAC;SAC9C,CAAC,CAAC,CAAC;QACJ,IAAI,MAAM,CAAC,QAAQ,EAAE,CAAC;YACpB,OAAO,CAAC,MAAM,CAAC,KAAK,CAAC,IAAI,CAAC,wBAAwB,EAAE,IAAI,EAAE,EAAE,EAAE,EAAE,MAAM,CAAC,QAAQ,EAAE,CAAC,CAAC,CAAC;QACtF,CAAC;QAED,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,SAAS,CAAC,MAAM,EAAE,IAAI,EAAE,CAAC,CAAC,CAAC,CAAC;IAC/C,CAAC;IAAC,OAAO,GAAY,EAAE,CAAC;QACtB,iDAAiD;QACjD,IAAI,GAAG,YAAY,OAAO;YAAE,MAAM,GAAG,CAAC;QACtC,OAAO,CAAC,KAAK,CAAC,IAAI,CAAC,yBAAyB,EAAE,IAAI,EAAE,EAAE,OAAO,EAAG,GAAa,CAAC,OAAO,EAAE,CAAC,CAAC,CAAC;QAC1F,MAAM,IAAI,OAAO,CAAC,CAAC,CAAC,CAAC;IACvB,CAAC;AACH,CAAC"}
@@ -0,0 +1,2 @@
1
+ export declare function execute(argv: string[]): Promise<void>;
2
+ //# sourceMappingURL=init.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"init.d.ts","sourceRoot":"","sources":["../../../../src/cli/commands/init.ts"],"names":[],"mappings":"AAsFA,wBAAsB,OAAO,CAAC,IAAI,EAAE,MAAM,EAAE,GAAG,OAAO,CAAC,IAAI,CAAC,CA4B3D"}
@@ -0,0 +1,110 @@
1
+ import { resolve, join } from 'node:path';
2
+ import { tCli, langFromArgv } from '../i18n.js';
3
+ import { COMMON_OPTIONS } from '../parse-run-config.js';
4
+ import { parseArgsStrictOrExit } from '../parse-strict.js';
5
+ const INIT_SAMPLES = `[
6
+ {
7
+ "sample_id": "s001",
8
+ "prompt": "审查以下代码",
9
+ "context": "function authenticate(username, password) {\\n const query = \`SELECT * FROM users WHERE name='\${username}' AND pass='\${password}'\`;\\n return db.execute(query);\\n}",
10
+ "rubric": "应识别 SQL 注入风险,建议使用参数化查询",
11
+ "assertions": [
12
+ { "type": "contains", "value": "SQL", "weight": 1 },
13
+ { "type": "contains", "value": "injection", "weight": 1 },
14
+ { "type": "regex", "pattern": "parameterized|prepared|placeholder|bind", "flags": "i", "weight": 0.5 },
15
+ { "type": "not_contains", "value": "looks good", "weight": 0.5 }
16
+ ],
17
+ "dimensions": {
18
+ "security": "是否准确识别出 SQL 注入漏洞并说明其危害",
19
+ "actionability": "是否给出可直接使用的参数化查询修复代码"
20
+ }
21
+ },
22
+ {
23
+ "sample_id": "s002",
24
+ "prompt": "审查以下代码",
25
+ "context": "async function fetchData(url) {\\n const res = await fetch(url);\\n const data = await res.json();\\n return data;\\n}",
26
+ "rubric": "应指出缺少错误处理(网络异常、非 JSON 响应、HTTP 错误状态码)",
27
+ "assertions": [
28
+ { "type": "contains", "value": "error handling", "weight": 1 },
29
+ { "type": "regex", "pattern": "try[\\\\s\\\\S]*catch|exception|error", "flags": "i", "weight": 1 },
30
+ { "type": "contains", "value": "status", "weight": 0.5 }
31
+ ],
32
+ "dimensions": {
33
+ "robustness": "是否指出了所有缺失的错误处理场景",
34
+ "actionability": "是否给出了完整的 try-catch 修复代码"
35
+ }
36
+ },
37
+ {
38
+ "sample_id": "s003",
39
+ "prompt": "审查以下代码",
40
+ "context": "function renderComment(comment) {\\n document.getElementById('output').innerHTML = '<p>' + comment + '</p>';\\n}",
41
+ "rubric": "应识别 XSS 风险,建议使用 textContent 或转义 HTML",
42
+ "assertions": [
43
+ { "type": "contains", "value": "XSS", "weight": 1 },
44
+ { "type": "regex", "pattern": "textContent|escape|sanitize|sanitizer", "flags": "i", "weight": 1 },
45
+ { "type": "contains", "value": "innerHTML", "weight": 0.5 }
46
+ ],
47
+ "dimensions": {
48
+ "security": "是否准确识别出 XSS 漏洞并说明攻击方式",
49
+ "actionability": "是否给出使用 textContent 或转义的修复代码"
50
+ }
51
+ }
52
+ ]
53
+ `;
54
+ // 模板带 Claude Code SKILL.md 兼容 frontmatter(name + description),让用户
55
+ // 可以把 init 出来的 SKILL.md 直接 deploy 到 ~/.claude/skills/ 给 Claude Code 用,
56
+ // 一份文件双向 dogfood(omk 评测 + Claude 部署)。omk 当前不 strip frontmatter,
57
+ // 它会跟着 leak 进 system prompt — 在 model 行为层面是无害噪声,跨 executor 一致。
58
+ const INIT_SKILL_V1 = `---
59
+ name: code-review-v1
60
+ description: 简单代码审查 skill,识别明显问题
61
+ ---
62
+
63
+ # Code review v1
64
+
65
+ 你是一个代码审查助手。请审查用户提供的代码,指出潜在问题。
66
+ `;
67
+ const INIT_SKILL_V2 = `---
68
+ name: code-review-v2
69
+ description: 多维度代码审查,覆盖安全 / 健壮 / 可维护 / 性能,带严重程度标注
70
+ ---
71
+
72
+ # Code review v2
73
+
74
+ 你是一个高级代码审查专家。请从以下维度审查用户提供的代码:
75
+
76
+ 1. 安全性:是否存在注入、XSS、敏感信息泄露等风险
77
+ 2. 健壮性:是否有适当的错误处理和边界检查
78
+ 3. 可维护性:命名是否清晰、结构是否合理
79
+ 4. 性能:是否存在明显的性能瓶颈
80
+
81
+ 对每个维度给出具体的改进建议,并标注严重程度(高/中/低)。
82
+ `;
83
+ export async function execute(argv) {
84
+ const lang = langFromArgv(argv);
85
+ // 走 helper 让未知 option fail-fast(e.g. `omk init --bogus`),
86
+ // 否则 argv[0] 直接当目录名, --bogus / --lang 都会被当成 dir 写文件。
87
+ const { positionals } = parseArgsStrictOrExit({
88
+ args: argv,
89
+ allowPositionals: true,
90
+ options: { ...COMMON_OPTIONS },
91
+ });
92
+ const targetDir = resolve(positionals[0] || '.');
93
+ const { writeFileSync, mkdirSync } = await import('node:fs');
94
+ // omk skill loader 把 `skills/<name>/SKILL.md` 子目录识别为 directory-skill,
95
+ // cwd 默认锚到 skill 根目录,后续可在同目录下放 assets / 子文档。
96
+ // 子目录主题化命名(code-review-v1 / code-review-v2)比泛 v1.md / v2.md 心智模型更清晰。
97
+ mkdirSync(join(targetDir, 'skills', 'code-review-v1'), { recursive: true });
98
+ mkdirSync(join(targetDir, 'skills', 'code-review-v2'), { recursive: true });
99
+ writeFileSync(join(targetDir, 'eval-samples.json'), INIT_SAMPLES);
100
+ writeFileSync(join(targetDir, 'skills', 'code-review-v1', 'SKILL.md'), INIT_SKILL_V1);
101
+ writeFileSync(join(targetDir, 'skills', 'code-review-v2', 'SKILL.md'), INIT_SKILL_V2);
102
+ console.log(tCli('cli.init.scaffolded', lang, { dir: targetDir }));
103
+ console.log('');
104
+ console.log(tCli('cli.init.next_steps_title', lang));
105
+ console.log(tCli('cli.init.next_step_edit_samples', lang));
106
+ console.log(tCli('cli.init.next_step_edit_skills', lang));
107
+ console.log(tCli('cli.init.next_step_run', lang));
108
+ console.log(tCli('cli.init.note_codex_executor', lang));
109
+ }
110
+ //# sourceMappingURL=init.js.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"init.js","sourceRoot":"","sources":["../../../../src/cli/commands/init.ts"],"names":[],"mappings":"AAAA,OAAO,EAAE,OAAO,EAAE,IAAI,EAAE,MAAM,WAAW,CAAC;AAC1C,OAAO,EAAE,IAAI,EAAE,YAAY,EAAE,MAAM,YAAY,CAAC;AAChD,OAAO,EAAE,cAAc,EAAE,MAAM,wBAAwB,CAAC;AACxD,OAAO,EAAE,qBAAqB,EAAE,MAAM,oBAAoB,CAAC;AAE3D,MAAM,YAAY,GAAG;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CAgDpB,CAAC;AAEF,kEAAkE;AAClE,uEAAuE;AACvE,gEAAgE;AAChE,+DAA+D;AAC/D,MAAM,aAAa,GAAG;;;;;;;;CAQrB,CAAC;AAEF,MAAM,aAAa,GAAG;;;;;;;;;;;;;;;CAerB,CAAC;AAEF,MAAM,CAAC,KAAK,UAAU,OAAO,CAAC,IAAc;IAC1C,MAAM,IAAI,GAAG,YAAY,CAAC,IAAI,CAAC,CAAC;IAChC,0DAA0D;IAC1D,qDAAqD;IACrD,MAAM,EAAE,WAAW,EAAE,GAAG,qBAAqB,CAAC;QAC5C,IAAI,EAAE,IAAI;QACV,gBAAgB,EAAE,IAAI;QACtB,OAAO,EAAE,EAAE,GAAG,cAAc,EAAE;KAC/B,CAAC,CAAC;IACH,MAAM,SAAS,GAAW,OAAO,CAAC,WAAW,CAAC,CAAC,CAAC,IAAI,GAAG,CAAC,CAAC;IACzD,MAAM,EAAE,aAAa,EAAE,SAAS,EAAE,GAAG,MAAM,MAAM,CAAC,SAAS,CAAC,CAAC;IAE7D,sEAAsE;IACtE,6CAA6C;IAC7C,qEAAqE;IACrE,SAAS,CAAC,IAAI,CAAC,SAAS,EAAE,QAAQ,EAAE,gBAAgB,CAAC,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;IAC5E,SAAS,CAAC,IAAI,CAAC,SAAS,EAAE,QAAQ,EAAE,gBAAgB,CAAC,EAAE,EAAE,SAAS,EAAE,IAAI,EAAE,CAAC,CAAC;IAC5E,aAAa,CAAC,IAAI,CAAC,SAAS,EAAE,mBAAmB,CAAC,EAAE,YAAY,CAAC,CAAC;IAClE,aAAa,CAAC,IAAI,CAAC,SAAS,EAAE,QAAQ,EAAE,gBAAgB,EAAE,UAAU,CAAC,EAAE,aAAa,CAAC,CAAC;IACtF,aAAa,CAAC,IAAI,CAAC,SAAS,EAAE,QAAQ,EAAE,gBAAgB,EAAE,UAAU,CAAC,EAAE,aAAa,CAAC,CAAC;IAEtF,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,qBAAqB,EAAE,IAAI,EAAE,EAAE,GAAG,EAAE,SAAS,EAAE,CAAC,CAAC,CAAC;IACnE,OAAO,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;IAChB,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,2BAA2B,EAAE,IAAI,CAAC,CAAC,CAAC;IACrD,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,iCAAiC,EAAE,IAAI,CAAC,CAAC,CAAC;IAC3D,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,gCAAgC,EAAE,IAAI,CAAC,CAAC,CAAC;IAC1D,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,wBAAwB,EAAE,IAAI,CAAC,CAAC,CAAC;IAClD,OAAO,CAAC,GAAG,CAAC,IAAI,CAAC,8BAA8B,EAAE,IAAI,CAAC,CAAC,CAAC;AAC1D,CAAC"}
@@ -0,0 +1,2 @@
1
+ export declare function execute(argv: string[]): Promise<void>;
2
+ //# sourceMappingURL=observe.d.ts.map
@@ -0,0 +1 @@
1
+ {"version":3,"file":"observe.d.ts","sourceRoot":"","sources":["../../../../src/cli/commands/observe.ts"],"names":[],"mappings":"AAOA,wBAAsB,OAAO,CAAC,IAAI,EAAE,MAAM,EAAE,GAAG,OAAO,CAAC,IAAI,CAAC,CA0E3D"}