oh-my-knowledge 0.33.0 → 0.35.0

package/dist/cli/commands/doctor.js

@@ -117,7 +117,7 @@ export default class Doctor extends BaseCommand {
         }),
         samples: Flags.string({
             description: bilingual({
-                zh: '样本文件路径（.json/.yaml）。不传则按 target / cwd 顺序自动发现。',
+                zh: '用例文件路径（.json/.yaml）。不传则按 target / cwd 顺序自动发现。',
                 en: 'Samples file path (.json/.yaml). Auto-detects from target / cwd if omitted.',
             }),
         }),
@@ -300,7 +300,7 @@ export function pruneDoctorHistory(dir, skillName, maxKeep) {
             continue;
         try {
             const data = JSON.parse(readFileSync(join(dir, file), 'utf-8'));
-            if (data?.kind !== 'doctor' || !Array.isArray(data.skills) || data.skills.length !== 1)
+            if (data?.reportKind !== 'doctor' || !Array.isArray(data.skills) || data.skills.length !== 1)
                 continue;
             if (data.skills[0].skillName !== skillName)
                 continue;
@@ -318,4 +318,3 @@ export function pruneDoctorHistory(dir, skillName, maxKeep) {
         catch { /* ignore */ }
     }
 }
-//# sourceMappingURL=doctor.js.map

package/dist/cli/commands/eval/gold/compare.d.ts

@@ -14,4 +14,3 @@ export default class EvalGoldCompare extends BaseCommand {
     };
     run(): Promise<void>;
 }
-//# sourceMappingURL=compare.d.ts.map

package/dist/cli/commands/eval/gold/compare.js

@@ -88,4 +88,3 @@ export default class EvalGoldCompare extends BaseCommand {
         });
     }
 }
-//# sourceMappingURL=compare.js.map

package/dist/cli/commands/eval/gold/index.d.ts

@@ -6,4 +6,3 @@ export default class EvalGold extends BaseCommand {
     };
     run(): Promise<void>;
 }
-//# sourceMappingURL=index.d.ts.map

package/dist/cli/commands/eval/gold/index.js

@@ -31,4 +31,3 @@ export default class EvalGold extends BaseCommand {
         this.exit(1);
     }
 }
-//# sourceMappingURL=index.js.map

package/dist/cli/commands/eval/gold/init.d.ts

@@ -8,4 +8,3 @@ export default class EvalGoldInit extends BaseCommand {
     };
     run(): Promise<void>;
 }
-//# sourceMappingURL=init.d.ts.map

package/dist/cli/commands/eval/gold/init.js

@@ -48,4 +48,3 @@ export default class EvalGoldInit extends BaseCommand {
         });
     }
 }
-//# sourceMappingURL=init.js.map

package/dist/cli/commands/eval/gold/validate.d.ts

@@ -9,4 +9,3 @@ export default class EvalGoldValidate extends BaseCommand {
     };
     run(): Promise<void>;
 }
-//# sourceMappingURL=validate.d.ts.map

package/dist/cli/commands/eval/gold/validate.js

@@ -43,4 +43,3 @@ export default class EvalGoldValidate extends BaseCommand {
         });
     }
 }
-//# sourceMappingURL=validate.js.map

package/dist/cli/commands/eval/index.d.ts

@@ -9,6 +9,8 @@ export default class Eval extends BaseCommand {
         lang: import("@oclif/core/interfaces").OptionFlag<string, import("@oclif/core/interfaces").CustomOptions>;
         control: import("@oclif/core/interfaces").OptionFlag<string | undefined, import("@oclif/core/interfaces").CustomOptions>;
         treatment: import("@oclif/core/interfaces").OptionFlag<string | undefined, import("@oclif/core/interfaces").CustomOptions>;
+        'control-cwd': import("@oclif/core/interfaces").OptionFlag<string | undefined, import("@oclif/core/interfaces").CustomOptions>;
+        'treatment-cwd': import("@oclif/core/interfaces").OptionFlag<string | undefined, import("@oclif/core/interfaces").CustomOptions>;
         config: import("@oclif/core/interfaces").OptionFlag<string | undefined, import("@oclif/core/interfaces").CustomOptions>;
         samples: import("@oclif/core/interfaces").OptionFlag<string | undefined, import("@oclif/core/interfaces").CustomOptions>;
         'skill-dir': import("@oclif/core/interfaces").OptionFlag<string | undefined, import("@oclif/core/interfaces").CustomOptions>;
@@ -48,7 +50,7 @@ export default class Eval extends BaseCommand {
         'trivial-diff': import("@oclif/core/interfaces").OptionFlag<string | undefined, import("@oclif/core/interfaces").CustomOptions>;
         'report-only': import("@oclif/core/interfaces").BooleanFlag<boolean>;
         'no-gate': import("@oclif/core/interfaces").BooleanFlag<boolean>;
+        'no-evidence': import("@oclif/core/interfaces").BooleanFlag<boolean>;
     };
     run(): Promise<void>;
 }
-//# sourceMappingURL=index.d.ts.map

package/dist/cli/commands/eval/index.js

@@ -37,21 +37,44 @@ function applyGateExitCode(code, values, lang) {
     process.stderr.write(tCli('cli.run.report_only_gate_skipped', lang));
     return 0;
 }
-async function emitEvaluationVerdict(report, values) {
+async function emitEvaluationVerdict(report, values, lang) {
     const { computeVerdict, formatVerdictText } = await import('../../../eval-core/verdict.js');
     const result = computeVerdict(report, verdictOptions(values));
     console.log(formatVerdictText(result, { verbose: true }));
+    await recordEvidenceSafely(report, result.level, values, lang);
     return verdictPasses(result.level, result.headline) ? 0 : 1;
 }
+/**
+ * 把本次评测写成证据追加进**已纳管**记录(让 install 过的 skill 走到 measurable)。
+ * 永不致命:管理是 eval 的旁路,写入失败 / 无匹配记录都不影响 verdict 与 exit code。
+ * `--no-evidence` 关闭。仅对实际写入的记录打一行提示(无匹配则全静默)。
+ */
+async function recordEvidenceSafely(report, verdict, values, lang) {
+    if (values['no-evidence'] === true)
+        return;
+    try {
+        const { recordEvalEvidence } = await import('../../../managed/index.js');
+        const written = recordEvalEvidence(report, verdict, new Date().toISOString());
+        for (const w of written) {
+            process.stderr.write(tCli(w.bound ? 'cli.run.evidence_recorded' : 'cli.run.evidence_recorded_unbound', lang, { name: w.name }));
+        }
+    }
+    catch {
+        // 证据写入是旁路,任何异常都不该让评测失败
+    }
+}
 function batchItemFallbackReport(batch, item) {
     return {
-        kind: 'evaluation',
+        reportKind: 'evaluation',
         id: item.reportId,
         meta: {
             ...batch.meta,
             variants: ['baseline', item.name],
             sampleCount: item.sampleCount,
             totalCostUSD: item.totalCostUSD,
+            // item.artifactHash 来自子报告(走 aggregateReport 的整树哈),故 fallback 与之一致标 schemaVersion 3,
+            // 避免「树哈 artifactHashes + 错位 schemaVersion」的错配。
+            schemaVersion: 3,
             artifactHashes: item.artifactHash ? { [item.name]: item.artifactHash } : {},
         },
         summary: item.summary,
@@ -65,7 +88,7 @@ async function loadBatchChildReports(batch, reportsDir, lang) {
     const reports = [];
     for (const item of batch.items) {
         const loaded = await store.get(item.reportId);
-        if (loaded?.kind === 'evaluation') {
+        if (loaded?.reportKind === 'evaluation') {
             reports.push(loaded);
         }
         else {
@@ -83,6 +106,11 @@ async function emitBatchVerdict(report, reportsDir, values, lang) {
         treatment: child.meta.variants[1] ?? child.id,
         verdict: computeVerdict(child, verdictOptions(values)),
     }));
+    // batch 每个子报告各自是一份独立 skill 的评测 → 各自写证据。
+    for (const child of childReports) {
+        const v = results.find((r) => r.id === child.id)?.verdict.level ?? 'SOLO';
+        await recordEvidenceSafely(child, v, values, lang);
+    }
     const passed = results.filter((r) => verdictPasses(r.verdict.level, r.verdict.headline)).length;
     const failed = results.length - passed;
     const status = lang === 'zh'
@@ -100,7 +128,7 @@ async function emitBatchVerdict(report, reportsDir, values, lang) {
 }
 async function announceSavedReport({ report, filePath, reportsDir, values, lang, }) {
     const tally = computeRunTally(report);
-    process.stderr.write(tCli(report.kind === 'batch-evaluation' ? 'cli.run.batch_complete' : 'cli.run.eval_complete', lang));
+    process.stderr.write(tCli(report.reportKind === 'batch-evaluation' ? 'cli.run.batch_complete' : 'cli.run.eval_complete', lang));
     process.stderr.write(tCli('cli.run.tally', lang, tally));
     process.stderr.write(tCli('cli.run.report_saved', lang, { path: filePath }));
     if (!values['no-serve'] && process.stdout.isTTY) {
@@ -255,7 +283,7 @@ async function runEval(_args, flags, lang) {
         if (filePath) {
             await announceSavedReport({ report, filePath, reportsDir: config.outputDir, values, lang });
         }
-        const exitCode = await emitEvaluationVerdict(report, values);
+        const exitCode = await emitEvaluationVerdict(report, values, lang);
         throw new CliExit(applyGateExitCode(exitCode, values, lang));
     }
     catch (err) {
@@ -290,12 +318,24 @@ export default class Eval extends BaseCommand {
         lang: LANG_FLAG,
         // ── 实验角色 ──
         control: Flags.string({
-            description: bilingual({ zh: 'control variant 表达式', en: 'Control variant expr' }),
+            description: bilingual({ zh: 'control variant 表达式（仅 artifact 身份）', en: 'Control variant expr (artifact identity only)' }),
         }),
         treatment: Flags.string({
             description: bilingual({
-                zh: 'treatment variant 列表，逗号分隔',
-                en: 'Treatment variants, comma-separated',
+                zh: 'treatment variant 列表，逗号分隔（仅 artifact 身份）',
+                en: 'Treatment variants, comma-separated (artifact identity only)',
+            }),
+        }),
+        'control-cwd': Flags.string({
+            description: bilingual({
+                zh: 'control 的 runtime context 目录',
+                en: 'Runtime context dir for control',
+            }),
+        }),
+        'treatment-cwd': Flags.string({
+            description: bilingual({
+                zh: 'treatment 的 runtime context 目录列表，逗号分隔、与 --treatment 按序对齐（空位 = 无 cwd）',
+                en: 'Runtime context dirs for treatments, comma-separated, index-aligned with --treatment (blank = none)',
             }),
         }),
         config: Flags.string({
@@ -303,7 +343,7 @@ export default class Eval extends BaseCommand {
         }),
         samples: Flags.string({
             description: bilingual({
-                zh: '样本文件路径。默认 eval-samples.json，也接受 .yaml/.yml；自动发现 --skill-dir 下的 <skill>/.omk/samples.json。',
+                zh: '用例文件路径。默认 eval-samples.json，也接受 .yaml/.yml；自动发现 --skill-dir 下的 <skill>/.omk/samples.json。',
                 en: 'Samples file path. Defaults to eval-samples.json (also .yaml/.yml); auto-discovers <skill>/.omk/samples.json under --skill-dir.',
             }),
         }),
@@ -344,7 +384,7 @@ export default class Eval extends BaseCommand {
             parse: integerStringParser('--concurrency', { min: 1 }),
         }),
         timeout: Flags.string({
-            description: bilingual({ zh: '单样本超时秒，默认 600', en: 'Per-sample timeout sec, default 600' }),
+            description: bilingual({ zh: '单用例超时秒，默认 600', en: 'Per-sample timeout sec, default 600' }),
             parse: numberStringParser('--timeout', { min: 1 }),
         }),
         batch: Flags.boolean({
@@ -454,6 +494,12 @@ export default class Eval extends BaseCommand {
         'no-gate': Flags.boolean({
             description: bilingual({ zh: '关 verdict gate', en: 'Disable verdict gate' }),
         }),
+        'no-evidence': Flags.boolean({
+            description: bilingual({
+                zh: '不把本次评测写成证据追加进受管记录(默认会为已 install 的 skill 自动写)。',
+                en: 'Do not append this run as evidence to managed records (auto-written for installed skills by default).',
+            }),
+        }),
     };
     async run() {
         const { args, flags } = await this.parse(Eval);
@@ -463,4 +509,3 @@ export default class Eval extends BaseCommand {
         });
     }
 }
-//# sourceMappingURL=index.js.map

package/dist/cli/commands/evolve.d.ts

@@ -32,7 +32,12 @@ export default class Evolve extends BaseCommand {
         'reuse-latest-eval': import("@oclif/core/interfaces").BooleanFlag<boolean>;
         'improve-mode': import("@oclif/core/interfaces").OptionFlag<string, import("@oclif/core/interfaces").CustomOptions>;
         'holdout-ratio': import("@oclif/core/interfaces").OptionFlag<string, import("@oclif/core/interfaces").CustomOptions>;
+        'no-significance-gate': import("@oclif/core/interfaces").BooleanFlag<boolean>;
+        'significance-alpha': import("@oclif/core/interfaces").OptionFlag<string, import("@oclif/core/interfaces").CustomOptions>;
+        'test-ratio': import("@oclif/core/interfaces").OptionFlag<string, import("@oclif/core/interfaces").CustomOptions>;
+        'edit-budget': import("@oclif/core/interfaces").OptionFlag<string, import("@oclif/core/interfaces").CustomOptions>;
+        'no-edit-budget': import("@oclif/core/interfaces").BooleanFlag<boolean>;
+        'no-reject-memory': import("@oclif/core/interfaces").BooleanFlag<boolean>;
     };
     run(): Promise<void>;
 }
-//# sourceMappingURL=evolve.d.ts.map

package/dist/cli/commands/evolve.js

@@ -26,6 +26,13 @@ export async function runEvolve(args, flags, lang) {
         console.error(tCli('cli.evolve.specify_skill_path', lang));
         throw new CliExit(1);
     }
+    // A locked test set needs a separate val set to decide on; --test-ratio alone is a no-op.
+    if ((Number(flags['test-ratio']) || 0) > 0 && (Number(flags['holdout-ratio']) || 0) === 0) {
+        console.error(lang === 'zh'
+            ? '--test-ratio 需要配合 --holdout-ratio 使用（test 集只在留出 val 集时才有意义）。'
+            : '--test-ratio requires --holdout-ratio (a locked test set only makes sense alongside a held-out val set).');
+        throw new CliExit(2);
+    }
     const { resolveSkillInput } = await import('../lib/resolve-skill-input.js');
     let resolvedInput;
     try {
@@ -69,9 +76,16 @@ export async function runEvolve(args, flags, lang) {
             sampleFixMaxAttempts: Math.max(1, Number(flags['sample-fix-max-attempts']) || 2),
             reuseLatestEval: flags['reuse-latest-eval'],
             holdoutRatio: Number(flags['holdout-ratio']) || 0,
+            significanceGate: !flags['no-significance-gate'],
+            // parser 已保证是合法数字串；用 Number() 直取，不用 `|| default`——否则
+            // `--edit-budget 0`（关预算）/ `--significance-alpha 0` 会被 falsy 吞成默认值。
+            significanceAlpha: Number(flags['significance-alpha']),
+            testRatio: Number(flags['test-ratio']),
+            editBudget: flags['no-edit-budget'] ? 0 : Number(flags['edit-budget']),
+            rejectMemory: !flags['no-reject-memory'],
             improveMode: flags['improve-mode'] === 'rewrite' ? 'rewrite' : 'agent',
             onProgress: makeOnProgress(lang),
-            onRoundProgress({ round, totalRounds: _totalRounds, phase, score, delta, accepted, costUSD, costReported, error }) {
+            onRoundProgress({ round, totalRounds: _totalRounds, phase, score, delta, accepted, costUSD, costReported, error, significant }) {
                 // costReported=false 时显示「—」而不是 $0.0000(executor 不报 cost,如 codex)。
                 const fmtRoundCost = (c, r) => r ? `$${c.toFixed(4)}` : '—';
                 if (phase === 'baseline') {
@@ -86,7 +100,9 @@ export async function runEvolve(args, flags, lang) {
                 }
                 else if (phase === 'done') {
                     const delta_ = delta >= 0 ? `+${delta.toFixed(2)}` : delta.toFixed(2);
-                    const status = accepted ? '✓ ACCEPT' : '✗ REJECT';
+                    // 门拒掉「均分上升但不显著」的候选时，补一句原因，否则 (+0.0x) ✗ REJECT 看着矛盾。
+                    const rejectNote = !accepted && significant === false ? tCli('cli.evolve.reject_not_significant', lang) : '';
+                    const status = accepted ? '✓ ACCEPT' : `✗ REJECT${rejectNote}`;
                     process.stderr.write(tCli('cli.evolve.round_done', lang, {
                         round, score: score.toFixed(2), delta: delta_, status, cost: fmtRoundCost(costUSD, costReported !== false),
                     }));
@@ -110,6 +126,17 @@ export async function runEvolve(args, flags, lang) {
                     train: result.holdout.trainCount, holdout: result.holdout.holdoutCount,
                 }));
         }
+        if (result.gate?.underpowered) {
+            process.stderr.write(tCli('cli.evolve.gate_underpowered', lang));
+        }
+        if (result.test?.disabled) {
+            process.stderr.write(tCli('cli.evolve.test_disabled', lang, { ratio: result.test.ratio }));
+        }
+        else if (result.test && typeof result.generalizationScore === 'number') {
+            process.stderr.write(tCli('cli.evolve.generalization', lang, {
+                count: result.test.count, score: result.generalizationScore.toFixed(2),
+            }));
+        }
         process.stderr.write(tCli('cli.evolve.best_path', lang, {
             best: result.bestSkillPath, target: resolve(skillPath),
         }));
@@ -174,7 +201,7 @@ export default class Evolve extends BaseCommand {
         }),
         samples: Flags.string({
             description: bilingual({
-                zh: '样本文件路径，默认 eval-samples.json',
+                zh: '用例文件路径，默认 eval-samples.json',
                 en: 'Samples file, default eval-samples.json',
             }),
             default: 'eval-samples.json',
@@ -206,7 +233,7 @@ export default class Evolve extends BaseCommand {
             parse: integerStringParser('--concurrency', { min: 1 }),
         }),
         timeout: Flags.string({
-            description: bilingual({ zh: '单样本超时秒，默认 600', en: 'Per-sample timeout sec, default 600' }),
+            description: bilingual({ zh: '单用例超时秒，默认 600', en: 'Per-sample timeout sec, default 600' }),
             default: '600',
             parse: numberStringParser('--timeout', { min: 1 }),
         }),
@@ -244,7 +271,7 @@ export default class Evolve extends BaseCommand {
         }),
         'stop-on-assertions-pass': Flags.boolean({
             description: bilingual({
-                zh: '普通样本断言全过时提前停止',
+                zh: '普通用例断言全过时提前停止',
                 en: 'Stop early when normal samples pass assertions',
             }),
             default: false,
@@ -287,6 +314,51 @@ export default class Evolve extends BaseCommand {
             default: '0',
             parse: numberStringParser('--holdout-ratio', { min: 0, max: 1 }),
         }),
+        'no-significance-gate': Flags.boolean({
+            description: bilingual({
+                zh: '关掉显著性接受门，退回「候选分高一点点就收」的点估计判定（默认门开：只收统计显著的提升）',
+                en: 'Disable the significance accept gate, reverting to point-estimate accept (default: gate on — accept only statistically significant gains)',
+            }),
+            default: false,
+        }),
+        'significance-alpha': Flags.string({
+            description: bilingual({
+                zh: '显著性门的 diff CI 显著性水平（默认 0.05 = 95% CI）',
+                en: 'Significance level for the accept gate diff CI (default 0.05 = 95% CI)',
+            }),
+            default: '0.05',
+            parse: numberStringParser('--significance-alpha', { min: 0, max: 1 }),
+        }),
+        'test-ratio': Flags.string({
+            description: bilingual({
+                zh: '锁定 test 集比例（0..1，默认 0=关），需配 --holdout-ratio。全程不参与选择，收尾读一次给无偏泛化分',
+                en: 'Locked test fraction (0..1, default 0=off); requires --holdout-ratio. Never used for selection; read once at the end for an unbiased generalization score',
+            }),
+            default: '0',
+            parse: numberStringParser('--test-ratio', { min: 0, max: 1 }),
+        }),
+        'edit-budget': Flags.string({
+            description: bilingual({
+                zh: '单轮最多改动的 skill 行占比（默认 0.2）。超预算的候选评测前直接判拒，省 eval 成本',
+                en: 'Max fraction of skill lines a round may change (default 0.2). Over-budget candidates are rejected before evaluation, saving eval cost',
+            }),
+            default: '0.2',
+            parse: numberStringParser('--edit-budget', { min: 0, max: 1 }),
+        }),
+        'no-edit-budget': Flags.boolean({
+            description: bilingual({
+                zh: '关掉 edit budget 约束（允许任意大小的单轮改动）',
+                en: 'Disable the edit budget (allow arbitrarily large single-round edits)',
+            }),
+            default: false,
+        }),
+        'no-reject-memory': Flags.boolean({
+            description: bilingual({
+                zh: '关掉 rejected-edit 记忆（不把被拒改法回灌下一轮 prompt）',
+                en: 'Disable rejected-edit memory (do not feed rejected edits back into the next prompt)',
+            }),
+            default: false,
+        }),
     };
     async run() {
         const { args, flags } = await this.parse(Evolve);
@@ -296,4 +368,3 @@ export default class Evolve extends BaseCommand {
         });
     }
 }
-//# sourceMappingURL=evolve.js.map

package/dist/cli/commands/init.d.ts

@@ -13,4 +13,3 @@ export default class Init extends BaseCommand {
     };
     run(): Promise<void>;
 }
-//# sourceMappingURL=init.d.ts.map

package/dist/cli/commands/init.js

@@ -149,4 +149,3 @@ export default class Init extends BaseCommand {
         });
     }
 }
-//# sourceMappingURL=init.js.map

package/dist/cli/commands/install.d.ts

@@ -0,0 +1,24 @@
+import { BaseCommand } from '../oclif/base-command.js';
+export default class Install extends BaseCommand {
+    static description: string;
+    static examples: {
+        description: string;
+        command: string;
+    }[];
+    static args: {
+        input: import("@oclif/core/interfaces").Arg<string, Record<string, unknown>>;
+    };
+    static flags: {
+        lang: import("@oclif/core/interfaces").OptionFlag<string, import("@oclif/core/interfaces").CustomOptions>;
+        kind: import("@oclif/core/interfaces").OptionFlag<string | undefined, import("@oclif/core/interfaces").CustomOptions>;
+        to: import("@oclif/core/interfaces").OptionFlag<string, import("@oclif/core/interfaces").CustomOptions>;
+        dest: import("@oclif/core/interfaces").OptionFlag<string | undefined, import("@oclif/core/interfaces").CustomOptions>;
+        force: import("@oclif/core/interfaces").BooleanFlag<boolean>;
+        'dry-run': import("@oclif/core/interfaces").BooleanFlag<boolean>;
+        'git-url': import("@oclif/core/interfaces").OptionFlag<string | undefined, import("@oclif/core/interfaces").CustomOptions>;
+        'git-ref': import("@oclif/core/interfaces").OptionFlag<string | undefined, import("@oclif/core/interfaces").CustomOptions>;
+    };
+    run(): Promise<void>;
+    private installBuiltinAgentSkill;
+    private installManagedSkill;
+}