superlab 0.1.18 → 0.1.20

package/bin/superlab.cjs

@@ -12,6 +12,8 @@ const {
 } = require("../lib/install.cjs");
 const {
   archiveContext,
+  collaboratorEvalIssues,
+  collaboratorMissionIssues,
   pruneContext,
   refreshContext,
 } = require("../lib/context.cjs");
@@ -788,6 +790,8 @@ function printDoctor(options) {
   const templateIssues = validatePaperTemplateRoot(options.targetDir, config);
   const dataDecisionIssues = validateDataDecisions(options.targetDir);
   const evalProtocolIssues = validateEvalProtocol(options.targetDir);
+  const missionContextIssues = collaboratorMissionIssues(options.targetDir);
+  const collaboratorProtocolIssues = collaboratorEvalIssues(options.targetDir);
   const rootIssues = validateProjectRoots(options.targetDir, config);
   const autoStatus = getAutoStatus({ targetDir: options.targetDir });
   const autoIssues = autoStatus.issues;
@@ -806,6 +810,8 @@ function printDoctor(options) {
     templateIssues.length > 0 ||
     dataDecisionIssues.length > 0 ||
     evalProtocolIssues.length > 0 ||
+    missionContextIssues.length > 0 ||
+    collaboratorProtocolIssues.length > 0 ||
     rootIssues.length > 0 ||
     autoIssues.length > 0
   ) {
@@ -820,6 +826,8 @@ function printDoctor(options) {
       templateIssues,
       dataDecisionIssues,
       evalProtocolIssues,
+      missionContextIssues,
+      collaboratorProtocolIssues,
       rootIssues,
       autoIssues
     );

package/lib/auto_runner.cjs

@@ -253,7 +253,7 @@ async function startAutoMode({ targetDir, now = new Date() }) {
   const maxFailures = parseInteger(mode.maxFailures, 0);
   const maxIterations = parseInteger(mode.maxIterations, 1);
   const requiredArtifact = resolveRequiredArtifact(targetDir, mode.requiredTerminalArtifact);
-  const frozenCoreSnapshot = snapshotFrozenCore(targetDir, mode.frozenCore);
+  let frozenCoreSnapshot = snapshotFrozenCore(targetDir, mode.frozenCore);
   const { loopStages, finalStages } = splitAutoStages(mode.allowedStages);
   const executedStages = [];
   let failureCount = 0;
@@ -363,13 +363,14 @@ async function startAutoMode({ targetDir, now = new Date() }) {
           nextRung,
           decision: rungId ? `completed rung ${rungId}` : `completed stage ${stage}`,
         });
-        refreshContext({ targetDir });
-
         const frozenCoreChanges = detectFrozenCoreChanges(frozenCoreSnapshot);
         if (frozenCoreChanges.length > 0) {
           failAutoMode(`frozen core changed: ${frozenCoreChanges.join(", ")}`);
         }
 
+        refreshContext({ targetDir });
+        frozenCoreSnapshot = snapshotFrozenCore(targetDir, mode.frozenCore);
+
         const stopCheck = await runCheckCommand({
           targetDir,
           label: `stop check after ${rungId || stage}`,
@@ -437,12 +438,13 @@ async function startAutoMode({ targetDir, now = new Date() }) {
       decision: `promotion policy matched after ${label}`,
     });
     promotionApplied = true;
-    refreshContext({ targetDir });
-    verifyPromotionWriteback(targetDir, promotionSnapshot);
     const frozenCoreChangesAfterPromotion = detectFrozenCoreChanges(frozenCoreSnapshot);
     if (frozenCoreChangesAfterPromotion.length > 0) {
       failAutoMode(`frozen core changed: ${frozenCoreChangesAfterPromotion.join(", ")}`);
     }
+    refreshContext({ targetDir });
+    verifyPromotionWriteback(targetDir, promotionSnapshot);
+    frozenCoreSnapshot = snapshotFrozenCore(targetDir, mode.frozenCore);
   };
 
   if (evalProtocol.experimentRungs.length > 0) {

package/lib/context.cjs

@@ -2,6 +2,69 @@ const fs = require("node:fs");
 const path = require("node:path");
 const { parseEvalProtocol } = require("./eval_protocol.cjs");
 
+const PLACEHOLDER_VALUES = new Set(["", "tbd", "none", "待补充", "无"]);
+const MISSION_COLLABORATOR_FIELDS = [
+  { name: "One-sentence problem", labels: ["One-sentence problem", "一句话问题"] },
+  { name: "Why it matters", labels: ["Why it matters", "为什么重要"] },
+  { name: "Primary metric", labels: ["Primary metric", "主指标"] },
+  { name: "Success threshold", labels: ["Success threshold", "成功阈值"] },
+  { name: "Dataset or benchmark scope", labels: ["Dataset or benchmark scope", "数据集或 benchmark 范围"] },
+  { name: "Approved direction", labels: ["Approved direction", "已批准方向"] },
+];
+const EVAL_COLLABORATOR_FIELDS = [
+  { name: "Primary evaluation objective", labels: ["Primary evaluation objective", "主评估目标"] },
+  { name: "Primary metrics", labels: ["Primary metrics", "主指标"] },
+  { name: "Secondary metrics", labels: ["Secondary metrics", "次级指标"] },
+  { name: "Table plan", labels: ["Table plan", "主表计划"] },
+  { name: "Metric glossary", labels: ["Metric glossary", "指标释义"] },
+  { name: "Background sources", labels: ["Background sources", "背景来源"] },
+  {
+    name: "Method and baseline source papers",
+    labels: ["Method and baseline source papers", "方法与基线来源论文"],
+  },
+  {
+    name: "Method and baseline implementation source",
+    labels: ["Method and baseline implementation source", "方法与基线实现来源"],
+  },
+  { name: "Metric source papers", labels: ["Metric source papers", "指标来源论文"] },
+  { name: "Required output artifacts", labels: ["Required output artifacts", "必要输出工件"] },
+];
+const REPORT_FIELDS = {
+  problem: ["Research problem in plain language", "研究问题白话解释", "研究问题"],
+  whyItMatters: ["Why this problem matters", "为什么这个问题重要"],
+  setting: ["What setting or workflow this report is actually about", "这份报告实际对应的场景或流程"],
+  primaryMetrics: ["Primary metrics", "主指标"],
+  secondaryMetrics: ["Secondary metrics", "次级指标"],
+  requiredTerminalEvidence: ["Required terminal evidence", "必要终局证据"],
+  metricGuidePrimary: ["Primary metric plain-language explanation", "主指标白话解释"],
+  metricGuideSecondary: ["Secondary metric plain-language explanation", "次级指标白话解释"],
+  metricGuideSupport: [
+    "Health or support metrics and why they are not the main claim",
+    "健康度或支持性指标以及它们为什么不是主 claim",
+  ],
+  backgroundSources: ["Most important background papers or benchmark references", "最重要的背景论文或 benchmark 参考"],
+  backgroundAnchors: ["Why these are the right background anchors", "为什么这些是合适的背景锚点"],
+  methodBasis: ["Our method source or implementation basis", "我们的方法来源或实现基础"],
+  baselineSourcePapers: ["Baseline and comparison source papers", "基线与对比方法来源论文"],
+  baselineImplementationSources: [
+    "Baseline and comparison implementation sources",
+    "基线与对比方法实现来源",
+  ],
+  metricSourcePapers: ["Metric source papers", "指标来源论文"],
+  metricImplementationSource: ["Metric implementation source", "指标实现来源"],
+  metricDeviation: ["Deviation from original implementation", "与原始实现的偏差"],
+  datasets: ["Datasets", "数据集"],
+  baselines: ["Baselines", "基线"],
+  metrics: ["Metrics", "指标"],
+  finalPerformanceSummary: ["Final performance summary", "最终表现总结"],
+  tableCoverage: ["Table coverage", "表格覆盖范围"],
+};
+const TERMINOLOGY_FIELDS = {
+  methodName: ["Method name", "方法名"],
+  shortName: ["Short name or acronym", "简称或缩写"],
+  contributionBullets: ["Contribution bullets", "贡献 bullets", "Contribution bullets："],
+};
+
 function contextFile(targetDir, name) {
   return path.join(targetDir, ".lab", "context", name);
 }
@@ -58,6 +121,278 @@ function joinNonEmpty(parts, separator = "; ") {
   return parts.filter(Boolean).join(separator);
 }
 
+function isMeaningful(value) {
+  return !PLACEHOLDER_VALUES.has((value || "").trim().toLowerCase());
+}
+
+function readWorkflowConfig(targetDir) {
+  const configPath = path.join(targetDir, ".lab", "config", "workflow.json");
+  if (!fs.existsSync(configPath)) {
+    return {};
+  }
+  try {
+    return JSON.parse(fs.readFileSync(configPath, "utf8"));
+  } catch {
+    return {};
+  }
+}
+
+function resolveProjectPath(targetDir, configuredPath) {
+  if (!configuredPath || typeof configuredPath !== "string") {
+    return "";
+  }
+  return path.resolve(targetDir, configuredPath);
+}
+
+function getCollaboratorDeliverablePaths(targetDir) {
+  const config = readWorkflowConfig(targetDir);
+  const deliverablesRoot = resolveProjectPath(targetDir, config.deliverables_root || "docs/research");
+  return {
+    deliverablesRoot,
+    reportPath: path.join(deliverablesRoot, "report.md"),
+    mainTablesPath: path.join(deliverablesRoot, "main-tables.md"),
+  };
+}
+
+function hasCollaboratorFacingDeliverables(targetDir) {
+  const { reportPath, mainTablesPath } = getCollaboratorDeliverablePaths(targetDir);
+  return fs.existsSync(reportPath) || fs.existsSync(mainTablesPath);
+}
+
+function missingCollaboratorFields(text, fields) {
+  return fields.filter((field) => !isMeaningful(extractValue(text, field.labels))).map((field) => field.name);
+}
+
+function collaboratorMissionIssues(targetDir) {
+  if (!hasCollaboratorFacingDeliverables(targetDir)) {
+    return [];
+  }
+  const mission = readFileIfExists(contextFile(targetDir, "mission.md"));
+  if (!mission) {
+    return [];
+  }
+  const missing = missingCollaboratorFields(mission, MISSION_COLLABORATOR_FIELDS);
+  return missing.length > 0
+    ? [`mission context is still skeletal for collaborator-facing reporting: ${missing.join(", ")}`]
+    : [];
+}
+
+function collaboratorEvalIssues(targetDir) {
+  if (!hasCollaboratorFacingDeliverables(targetDir)) {
+    return [];
+  }
+  const protocol = readFileIfExists(contextFile(targetDir, "eval-protocol.md"));
+  if (!protocol) {
+    return [];
+  }
+  const missing = missingCollaboratorFields(protocol, EVAL_COLLABORATOR_FIELDS);
+  return missing.length > 0
+    ? [`evaluation protocol is still skeletal for collaborator-facing reporting: ${missing.join(", ")}`]
+    : [];
+}
+
+function extractReportValue(reportText, key) {
+  return extractValue(reportText, REPORT_FIELDS[key] || []);
+}
+
+function mergePreferred(existingValue, ...candidates) {
+  if (isMeaningful(existingValue)) {
+    return existingValue;
+  }
+  for (const candidate of candidates) {
+    if (isMeaningful(candidate)) {
+      return candidate;
+    }
+  }
+  return "";
+}
+
+function firstMetric(metrics) {
+  return (metrics || "")
+    .split(/[;,]/)
+    .map((value) => value.trim())
+    .filter(Boolean)[0] || "";
+}
+
+function buildMissionContextText(lang, fields) {
+  if (lang === "zh") {
+    return `# 研究主线
+
+## 核心问题
+
+- 一句话问题: ${fields.problem || "待补充"}
+- 为什么重要: ${fields.whyItMatters || "待补充"}
+- 目标失败场景: ${fields.targetFailureCase || "待补充"}
+
+## 成功标准
+
+- 主指标: ${fields.primaryMetric || "待补充"}
+- 成功阈值: ${fields.successThreshold || "待补充"}
+- 必须对比的 baseline: ${fields.requiredBaselineComparison || "待补充"}
+- 最小证据要求: ${fields.minimumEvidenceRequirement || "待补充"}
+
+## 冻结边界
+
+- 数据集或 benchmark 范围: ${fields.datasetScope || "待补充"}
+- 切分策略: ${fields.splitPolicy || "待补充"}
+- 评估协议: ${fields.evaluationProtocol || "待补充"}
+- 硬约束: ${fields.hardConstraints || "待补充"}
+
+## 当前状态
+
+- 已批准方向: ${fields.approvedDirection || "待补充"}
+- 当前 owner 或会话: ${fields.currentOwner || "待补充"}
+- 最近一次允许更新 mission 的阶段: ${fields.latestStage || "待补充"}
+- 回填来源: ${fields.hydrationProvenance || "待补充"}
+- 协作者可读状态: ${fields.collaboratorReadyStatus || "待补充"}
+`;
+  }
+
+  return `# Research Mission
+
+## Core Problem
+
+- One-sentence problem: ${fields.problem || "TBD"}
+- Why it matters: ${fields.whyItMatters || "TBD"}
+- Target failure case: ${fields.targetFailureCase || "TBD"}
+
+## Success Criteria
+
+- Primary metric: ${fields.primaryMetric || "TBD"}
+- Success threshold: ${fields.successThreshold || "TBD"}
+- Required baseline comparison: ${fields.requiredBaselineComparison || "TBD"}
+- Minimum evidence requirement: ${fields.minimumEvidenceRequirement || "TBD"}
+
+## Frozen Boundaries
+
+- Dataset or benchmark scope: ${fields.datasetScope || "TBD"}
+- Split policy: ${fields.splitPolicy || "TBD"}
+- Evaluation protocol: ${fields.evaluationProtocol || "TBD"}
+- Hard constraints: ${fields.hardConstraints || "TBD"}
+
+## Current Status
+
+- Approved direction: ${fields.approvedDirection || "TBD"}
+- Current owner or session: ${fields.currentOwner || "TBD"}
+- Latest stage to update this mission: ${fields.latestStage || "TBD"}
+- Hydration provenance: ${fields.hydrationProvenance || "TBD"}
+- Collaborator-ready status: ${fields.collaboratorReadyStatus || "TBD"}
+`;
+}
+
+function buildEvalProtocolText(lang, fields, rungs) {
+  const rungBlocks = (rungs || [])
+    .map((rung) => {
+      if (lang === "zh") {
+        return `### Rung: ${rung.id}
+
+- 阶段: ${rung.stage || "待补充"}
+- 目标: ${rung.goal || "待补充"}
+- 命令: ${rung.command || "待补充"}
+- 监视目标: ${rung.watch || "待补充"}
+- gate 命令: ${rung.gate || "待补充"}
+- 通过后: ${rung.onPass || "待补充"}
+- 失败后: ${rung.onFail || "待补充"}
+- 停止后: ${rung.onStop || "待补充"}`;
+      }
+      return `### Rung: ${rung.id}
+
+- Stage: ${rung.stage || "TBD"}
+- Goal: ${rung.goal || "TBD"}
+- Command: ${rung.command || "TBD"}
+- Watch: ${rung.watch || "TBD"}
+- Gate: ${rung.gate || "TBD"}
+- On pass: ${rung.onPass || "TBD"}
+- On fail: ${rung.onFail || "TBD"}
+- On stop: ${rung.onStop || "TBD"}`;
+    })
+    .join("\n\n");
+
+  if (lang === "zh") {
+    return `# 评估协议
+
+用这份文件定义 \`/lab:run\`、\`/lab:iterate\`、\`/lab:auto\` 和 \`/lab:report\` 共用的论文导向评估目标、主表计划、gate 与 benchmark ladder。
+
+## 主评估目标
+
+- 主评估目标: ${fields.primaryEvaluationObjective || "待补充"}
+- 主指标: ${fields.primaryMetrics || "待补充"}
+- 次级指标: ${fields.secondaryMetrics || "待补充"}
+- 必要终局证据: ${fields.requiredTerminalEvidence || "待补充"}
+
+## 主表计划
+
+- 主表计划: ${fields.tablePlan || "待补充"}
+- 每张表必须支撑的 claims: ${fields.requiredClaimsPerTable || "待补充"}
+
+## 指标释义
+
+- 指标释义: ${fields.metricGlossary || "待补充"}
+- 回填来源: ${fields.hydrationProvenance || "待补充"}
+- 背景来源: ${fields.backgroundSources || "待补充"}
+- 方法与基线来源论文: ${fields.methodAndBaselineSourcePapers || "待补充"}
+- 方法与基线实现来源: ${fields.methodAndBaselineImplementationSource || "待补充"}
+- 指标来源论文: ${fields.metricSourcePapers || "待补充"}
+- 指标实现来源: ${fields.metricImplementationSource || "待补充"}
+- 对比方法来源论文: ${fields.comparisonSourcePapers || "待补充"}
+- 对比方法实现来源: ${fields.comparisonImplementationSource || "待补充"}
+- 与原始实现的偏差: ${fields.deviationFromOriginalImplementation || "待补充"}
+
+## Gate Ladder
+
+- 实验阶梯: ${fields.experimentLadder || "待补充"}
+- benchmark 阶梯: ${fields.benchmarkLadder || "待补充"}
+- 对比方法 gate: ${fields.comparisonGate || "待补充"}
+- 升格 gate: ${fields.promotionGate || "待补充"}
+- 最小样本量: ${fields.minimumSampleSizes || "待补充"}
+- 必要输出工件: ${fields.requiredOutputArtifacts || "待补充"}
+
+${rungBlocks}
+`;
+  }
+
+  return `# Evaluation Protocol
+
+Use this file to define the paper-facing evaluation target, table plan, gates, and benchmark ladder shared by \`/lab:run\`, \`/lab:iterate\`, \`/lab:auto\`, and \`/lab:report\`.
+
+## Primary Evaluation Objective
+
+- Primary evaluation objective: ${fields.primaryEvaluationObjective || "TBD"}
+- Primary metrics: ${fields.primaryMetrics || "TBD"}
+- Secondary metrics: ${fields.secondaryMetrics || "TBD"}
+- Required terminal evidence: ${fields.requiredTerminalEvidence || "TBD"}
+
+## Table Plan
+
+- Table plan: ${fields.tablePlan || "TBD"}
+- Required claims per table: ${fields.requiredClaimsPerTable || "TBD"}
+
+## Metric Glossary
+
+- Metric glossary: ${fields.metricGlossary || "TBD"}
+- Hydration provenance: ${fields.hydrationProvenance || "TBD"}
+- Background sources: ${fields.backgroundSources || "TBD"}
+- Method and baseline source papers: ${fields.methodAndBaselineSourcePapers || "TBD"}
+- Method and baseline implementation source: ${fields.methodAndBaselineImplementationSource || "TBD"}
+- Metric source papers: ${fields.metricSourcePapers || "TBD"}
+- Metric implementation source: ${fields.metricImplementationSource || "TBD"}
+- Comparison source papers: ${fields.comparisonSourcePapers || "TBD"}
+- Comparison implementation source: ${fields.comparisonImplementationSource || "TBD"}
+- Deviation from original implementation: ${fields.deviationFromOriginalImplementation || "TBD"}
+
+## Gate Ladder
+
+- Experiment ladder: ${fields.experimentLadder || "TBD"}
+- Benchmark ladder: ${fields.benchmarkLadder || "TBD"}
+- Comparison gate: ${fields.comparisonGate || "TBD"}
+- Promotion gate: ${fields.promotionGate || "TBD"}
+- Minimum sample sizes: ${fields.minimumSampleSizes || "TBD"}
+- Required output artifacts: ${fields.requiredOutputArtifacts || "TBD"}
+
+${rungBlocks}
+`;
+}
+
 function extractClaim(text) {
   const blocks = text
     .split(/\n(?=\d+\.\s)/)
@@ -76,6 +411,264 @@ function labelValue(text, englishLabels, chineseLabels = []) {
   return extractValue(text, [...englishLabels, ...chineseLabels]);
 }
 
+function collectHydrationSources(targetDir) {
+  const { reportPath, mainTablesPath } = getCollaboratorDeliverablePaths(targetDir);
+  return [
+    fs.existsSync(reportPath) ? path.relative(targetDir, reportPath) : "",
+    fs.existsSync(mainTablesPath) ? path.relative(targetDir, mainTablesPath) : "",
+    readFileIfExists(contextFile(targetDir, "data-decisions.md")) ? ".lab/context/data-decisions.md" : "",
+    readFileIfExists(contextFile(targetDir, "state.md")) ? ".lab/context/state.md" : "",
+    readFileIfExists(contextFile(targetDir, "evidence-index.md")) ? ".lab/context/evidence-index.md" : "",
+  ].filter(Boolean);
+}
+
+function hydrateMissionContext(targetDir) {
+  if (!hasCollaboratorFacingDeliverables(targetDir)) {
+    return false;
+  }
+
+  const lang = readWorkflowLanguage(targetDir);
+  const missionText = readFileIfExists(contextFile(targetDir, "mission.md"));
+  const stateText = readFileIfExists(contextFile(targetDir, "state.md"));
+  const evidenceText = readFileIfExists(contextFile(targetDir, "evidence-index.md"));
+  const dataDecisions = readFileIfExists(contextFile(targetDir, "data-decisions.md"));
+  const reportText = readFileIfExists(getCollaboratorDeliverablePaths(targetDir).reportPath);
+  const evalProtocol = parseEvalProtocol(targetDir);
+  const hydrationSources = collectHydrationSources(targetDir);
+
+  const fields = {
+    problem: mergePreferred(
+      extractValue(missionText, ["One-sentence problem", "一句话问题"]),
+      extractReportValue(reportText, "problem"),
+      extractValue(stateText, ["Current objective", "当前目标", "Current objective"])
+    ),
+    whyItMatters: mergePreferred(
+      extractValue(missionText, ["Why it matters", "为什么重要"]),
+      extractReportValue(reportText, "whyItMatters")
+    ),
+    targetFailureCase: extractValue(missionText, ["Target failure case", "目标失败场景"]),
+    primaryMetric: mergePreferred(
+      extractValue(missionText, ["Primary metric", "主指标"]),
+      evalProtocol.primaryMetrics,
+      extractReportValue(reportText, "primaryMetrics"),
+      firstMetric(evalProtocol.primaryMetrics)
+    ),
+    successThreshold: mergePreferred(
+      extractValue(missionText, ["Success threshold", "成功阈值"]),
+      extractValue(readFileIfExists(contextFile(targetDir, "auto-outcome.md")), ["Terminal goal target", "终止目标目标值"])
+    ),
+    requiredBaselineComparison: extractValue(
+      missionText,
+      ["Required baseline comparison", "必须对比的 baseline"]
+    ),
+    minimumEvidenceRequirement: extractValue(
+      missionText,
+      ["Minimum evidence requirement", "最小证据要求"]
+    ),
+    datasetScope: mergePreferred(
+      extractValue(missionText, ["Dataset or benchmark scope", "数据集或 benchmark 范围"]),
+      extractValue(dataDecisions, ["Approved dataset package", "Approved datasets", "已批准数据集包", "已批准数据集"]),
+      extractReportValue(reportText, "datasets")
+    ),
+    splitPolicy: extractValue(missionText, ["Split policy", "切分策略"]),
+    evaluationProtocol: mergePreferred(
+      extractValue(missionText, ["Evaluation protocol", "评估协议"]),
+      ".lab/context/eval-protocol.md"
+    ),
+    hardConstraints: extractValue(missionText, ["Hard constraints", "硬约束"]),
+    approvedDirection: mergePreferred(
+      extractValue(missionText, ["Approved direction", "已批准方向"]),
+      extractClaim(evidenceText)
+    ),
+    currentOwner: extractValue(missionText, ["Current owner or session", "当前 owner 或会话"]),
+    latestStage: mergePreferred(
+      extractValue(missionText, ["Latest stage to update this mission", "最近一次允许更新 mission 的阶段"]),
+      extractValue(stateText, ["Active stage", "当前阶段", "Stage"])
+    ),
+  };
+
+  const before = missionText;
+  fields.hydrationProvenance = hydrationSources.length > 0 ? hydrationSources.join("; ") : "";
+  fields.collaboratorReadyStatus =
+    missingCollaboratorFields(buildMissionContextText(lang, fields), MISSION_COLLABORATOR_FIELDS).length === 0
+      ? "hydrated"
+      : "artifact-anchored interim";
+
+  const nextText = buildMissionContextText(lang, fields);
+  if (nextText !== before) {
+    writeContextFile(targetDir, "mission.md", nextText);
+    return true;
+  }
+  return false;
+}
+
+function hydrateEvalProtocol(targetDir) {
+  if (!hasCollaboratorFacingDeliverables(targetDir)) {
+    return false;
+  }
+
+  const lang = readWorkflowLanguage(targetDir);
+  const protocol = parseEvalProtocol(targetDir);
+  const missionText = readFileIfExists(contextFile(targetDir, "mission.md"));
+  const reportText = readFileIfExists(getCollaboratorDeliverablePaths(targetDir).reportPath);
+  const dataDecisions = readFileIfExists(contextFile(targetDir, "data-decisions.md"));
+  const hydrationSources = collectHydrationSources(targetDir);
+  const { reportPath, mainTablesPath } = getCollaboratorDeliverablePaths(targetDir);
+  const requiredOutputs = [
+    fs.existsSync(reportPath) ? path.relative(targetDir, reportPath) : "",
+    fs.existsSync(mainTablesPath) ? path.relative(targetDir, mainTablesPath) : "",
+    ".lab/context/evidence-index.md",
+  ]
+    .filter(Boolean)
+    .join(", ");
+
+  const fields = {
+    primaryEvaluationObjective: mergePreferred(
+      protocol.primaryEvaluationObjective,
+      extractValue(missionText, ["One-sentence problem", "一句话问题"]),
+      extractReportValue(reportText, "problem")
+    ),
+    primaryMetrics: mergePreferred(
+      protocol.primaryMetrics,
+      extractValue(missionText, ["Primary metric", "主指标"]),
+      extractReportValue(reportText, "primaryMetrics")
+    ),
+    secondaryMetrics: mergePreferred(
+      protocol.secondaryMetrics,
+      extractReportValue(reportText, "secondaryMetrics")
+    ),
+    requiredTerminalEvidence: mergePreferred(
+      protocol.requiredTerminalEvidence,
+      extractReportValue(reportText, "requiredTerminalEvidence"),
+      requiredOutputs
+    ),
+    tablePlan: mergePreferred(
+      protocol.tablePlan,
+      fs.existsSync(mainTablesPath) ? `See ${path.relative(targetDir, mainTablesPath)}` : ""
+    ),
+    requiredClaimsPerTable: protocol.requiredClaimsPerTable,
+    metricGlossary: mergePreferred(
+      protocol.metricGlossary,
+      joinNonEmpty(
+        [
+          extractReportValue(reportText, "metricGuidePrimary"),
+          extractReportValue(reportText, "metricGuideSecondary"),
+          extractReportValue(reportText, "metricGuideSupport"),
+        ],
+        " | "
+      )
+    ),
+    backgroundSources: mergePreferred(
+      protocol.backgroundSources,
+      extractReportValue(reportText, "backgroundSources"),
+      extractValue(dataDecisions, ["Papers that used the approved datasets", "使用过已批准数据集的论文", "使用过该数据集的论文"])
+    ),
+    methodAndBaselineSourcePapers: mergePreferred(
+      protocol.methodAndBaselineSourcePapers,
+      extractReportValue(reportText, "baselineSourcePapers")
+    ),
+    methodAndBaselineImplementationSource: mergePreferred(
+      protocol.methodAndBaselineImplementationSource,
+      extractReportValue(reportText, "baselineImplementationSources")
+    ),
+    metricSourcePapers: mergePreferred(
+      protocol.metricSourcePapers,
+      extractReportValue(reportText, "metricSourcePapers")
+    ),
+    metricImplementationSource: mergePreferred(
+      protocol.metricImplementationSource,
+      extractReportValue(reportText, "metricImplementationSource")
+    ),
+    comparisonSourcePapers: mergePreferred(
+      protocol.comparisonSourcePapers,
+      extractReportValue(reportText, "baselineSourcePapers")
+    ),
+    comparisonImplementationSource: mergePreferred(
+      protocol.comparisonImplementationSource,
+      extractReportValue(reportText, "baselineImplementationSources")
+    ),
+    deviationFromOriginalImplementation: mergePreferred(
+      protocol.deviationFromOriginalImplementation,
+      extractReportValue(reportText, "metricDeviation")
+    ),
+    benchmarkLadder: protocol.benchmarkLadder,
+    experimentLadder: protocol.experimentLadder,
+    comparisonGate: protocol.comparisonGate,
+    promotionGate: protocol.promotionGate,
+    minimumSampleSizes: protocol.minimumSampleSizes,
+    requiredOutputArtifacts: mergePreferred(protocol.requiredOutputArtifacts, requiredOutputs),
+    hydrationProvenance: hydrationSources.length > 0 ? hydrationSources.join("; ") : "",
+  };
+
+  const nextText = buildEvalProtocolText(lang, fields, protocol.experimentRungs);
+  if (nextText !== protocol.text) {
+    writeContextFile(targetDir, "eval-protocol.md", nextText);
+    return true;
+  }
+  return false;
+}
+
+function getCollaboratorReportStatus(targetDir) {
+  const missionIssues = collaboratorMissionIssues(targetDir);
+  const evalIssues = collaboratorEvalIssues(targetDir);
+  const issues = missionIssues.concat(evalIssues);
+  if (issues.length > 0) {
+    return {
+      mode: "artifact-anchored interim",
+      readiness: "hydrated but incomplete",
+      reason: issues.join(" | "),
+      issues,
+    };
+  }
+  return {
+    mode: "collaborator-ready",
+    readiness: "ready",
+    reason: "canonical mission and evaluation context are complete enough for collaborator-facing reporting",
+    issues: [],
+  };
+}
+
+function upsertSection(text, heading, bodyLines) {
+  const sectionText = `${heading}\n\n${bodyLines.join("\n")}`.trimEnd();
+  const pattern = new RegExp(`^${heading.replace(/[.*+?^${}()|[\]\\]/g, "\\$&")}\\s*$[\\s\\S]*?(?=^##\\s|\\Z)`, "m");
+  if (pattern.test(text)) {
+    return text.replace(pattern, sectionText);
+  }
+  return `${text.trimEnd()}\n\n${sectionText}\n`;
+}
+
+function syncCollaboratorReportStatus(targetDir, status) {
+  const { reportPath } = getCollaboratorDeliverablePaths(targetDir);
+  if (!fs.existsSync(reportPath)) {
+    return false;
+  }
+  const reportText = fs.readFileSync(reportPath, "utf8");
+  const heading = /^#\s/.test(reportText) ? "## Report Status" : "# Report Status";
+  const nextText = upsertSection(reportText, heading, [
+    `- Report mode: ${status.mode}`,
+    `- Canonical context readiness: ${status.readiness}`,
+    `- Why the current mode is appropriate: ${status.reason}`,
+  ]);
+  if (nextText !== reportText) {
+    fs.writeFileSync(reportPath, nextText.trimEnd() + "\n");
+    return true;
+  }
+  return false;
+}
+
+function hydrateCanonicalContext(targetDir) {
+  const missionChanged = hydrateMissionContext(targetDir);
+  const evalChanged = hydrateEvalProtocol(targetDir);
+  const reportStatus = getCollaboratorReportStatus(targetDir);
+  const reportChanged = syncCollaboratorReportStatus(targetDir, reportStatus);
+  return {
+    missionChanged,
+    evalChanged,
+    reportChanged,
+    reportStatus,
+  };
+}
+
 function renderSummary(lang, data) {
   if (lang === "zh") {
     return `# 研究摘要
@@ -95,6 +688,11 @@ function renderSummary(lang, data) {
 - Auto stop reason: ${data.autoStopReason || "待补充"}
 - Auto final artifact: ${data.autoFinalArtifact || "待补充"}
 - Auto final rung: ${data.autoFinalRung || "待补充"}
+- Collaborator report mode: ${data.reportMode || "待补充"}
+- Canonical context readiness: ${data.reportReadiness || "待补充"}
+- Why this report mode is active: ${data.reportModeReason || "待补充"}
+- Method name: ${data.methodName || "待补充"}
+- Contribution bullets: ${data.contributionBullets || "待补充"}
 - Eval objective: ${data.evalObjective || "待补充"}
 - Primary metrics: ${data.evalPrimaryMetrics || "待补充"}
 - Secondary metrics: ${data.evalSecondaryMetrics || "待补充"}
@@ -153,6 +751,11 @@ function renderSummary(lang, data) {
 - Auto stop reason: ${data.autoStopReason || "TBD"}
 - Auto final artifact: ${data.autoFinalArtifact || "TBD"}
 - Auto final rung: ${data.autoFinalRung || "TBD"}
+- Collaborator report mode: ${data.reportMode || "TBD"}
+- Canonical context readiness: ${data.reportReadiness || "TBD"}
+- Why this report mode is active: ${data.reportModeReason || "TBD"}
+- Method name: ${data.methodName || "TBD"}
+- Contribution bullets: ${data.contributionBullets || "TBD"}
 - Eval objective: ${data.evalObjective || "TBD"}
 - Primary metrics: ${data.evalPrimaryMetrics || "TBD"}
 - Secondary metrics: ${data.evalSecondaryMetrics || "TBD"}
@@ -266,6 +869,11 @@ ${data.problem || "待补充"}
 - Auto stop reason: ${data.autoStopReason || "待补充"}
 - Auto final artifact: ${data.autoFinalArtifact || "待补充"}
 - Auto final rung: ${data.autoFinalRung || "待补充"}
+- Collaborator report mode: ${data.reportMode || "待补充"}
+- Canonical context readiness: ${data.reportReadiness || "待补充"}
+- Why this report mode is active: ${data.reportModeReason || "待补充"}
+- Method name: ${data.methodName || "待补充"}
+- Contribution bullets: ${data.contributionBullets || "待补充"}
 - Eval objective: ${data.evalObjective || "待补充"}
 - Primary metrics: ${data.evalPrimaryMetrics || "待补充"}
 - Secondary metrics: ${data.evalSecondaryMetrics || "待补充"}
@@ -335,6 +943,11 @@ ${data.problem || "TBD"}
 - Auto stop reason: ${data.autoStopReason || "TBD"}
 - Auto final artifact: ${data.autoFinalArtifact || "TBD"}
 - Auto final rung: ${data.autoFinalRung || "TBD"}
+- Collaborator report mode: ${data.reportMode || "TBD"}
+- Canonical context readiness: ${data.reportReadiness || "TBD"}
+- Why this report mode is active: ${data.reportModeReason || "TBD"}
+- Method name: ${data.methodName || "TBD"}
+- Contribution bullets: ${data.contributionBullets || "TBD"}
 - Eval objective: ${data.evalObjective || "TBD"}
 - Primary metrics: ${data.evalPrimaryMetrics || "TBD"}
 - Secondary metrics: ${data.evalSecondaryMetrics || "TBD"}
@@ -381,11 +994,13 @@ ${data.problem || "TBD"}
 }
 
 function buildContextSnapshot(targetDir) {
+  const reportStatus = getCollaboratorReportStatus(targetDir);
   const mission = readFileIfExists(contextFile(targetDir, "mission.md"));
   const state = readFileIfExists(contextFile(targetDir, "state.md"));
   const evidence = readFileIfExists(contextFile(targetDir, "evidence-index.md"));
   const questions = readFileIfExists(contextFile(targetDir, "open-questions.md"));
   const dataDecisions = readFileIfExists(contextFile(targetDir, "data-decisions.md"));
+  const terminologyLock = readFileIfExists(contextFile(targetDir, "terminology-lock.md"));
   const autoMode = readFileIfExists(contextFile(targetDir, "auto-mode.md"));
   const autoStatus = readFileIfExists(contextFile(targetDir, "auto-status.md"));
   const autoOutcome = readFileIfExists(contextFile(targetDir, "auto-outcome.md"));
@@ -592,6 +1207,11 @@ function buildContextSnapshot(targetDir) {
     autoStopReason: extractValue(autoOutcome, ["Stop reason", "停止原因"]),
     autoFinalArtifact: extractValue(autoOutcome, ["Final artifact", "最终工件"]),
     autoFinalRung: extractValue(autoOutcome, ["Final rung", "最终 rung"]),
+    reportMode: reportStatus.mode,
+    reportReadiness: reportStatus.readiness,
+    reportModeReason: reportStatus.reason,
+    methodName: extractValue(terminologyLock, TERMINOLOGY_FIELDS.methodName),
+    contributionBullets: extractValue(terminologyLock, TERMINOLOGY_FIELDS.contributionBullets),
     evalObjective: evalProtocol.primaryEvaluationObjective,
     evalPrimaryMetrics: evalProtocol.primaryMetrics,
     evalSecondaryMetrics: evalProtocol.secondaryMetrics,
@@ -619,6 +1239,7 @@ function writeContextFile(targetDir, name, content) {
 }
 
 function refreshContext({ targetDir }) {
+  hydrateCanonicalContext(targetDir);
   const lang = readWorkflowLanguage(targetDir);
   const snapshot = buildContextSnapshot(targetDir);
   writeContextFile(targetDir, "summary.md", renderSummary(lang, snapshot));
@@ -695,6 +1316,11 @@ function archiveContext({ targetDir, now = new Date() }) {
 
 module.exports = {
   archiveContext,
+  collaboratorEvalIssues,
+  collaboratorMissionIssues,
+  getCollaboratorReportStatus,
+  hasCollaboratorFacingDeliverables,
+  hydrateCanonicalContext,
   pruneContext,
   refreshContext,
 };

package/lib/i18n.cjs

@@ -289,7 +289,11 @@ const ZH_SKILL_FILES = {
 
 ## 必要输出
 
+- 报告状态：collaborator-ready 或 artifact-anchored interim
 - 给用户看的总结
+- 问题与背景的白话说明
+- 数据集场景说明
+- 贡献总结
 - 方法概述
 - 选定指标摘要
 - 指标白话释义
@@ -301,6 +305,7 @@ const ZH_SKILL_FILES = {
 - 失败尝试
 - 局限性
 - 下一步
+- 单独列出的工件状态，而不是混进已验证结果
 
 ## 上下文读取
 
@@ -312,6 +317,8 @@ const ZH_SKILL_FILES = {
 
 ## 上下文写回
 
+- \`.lab/context/mission.md\`
+- \`.lab/context/eval-protocol.md\`
 - \`.lab/context/state.md\`
 - \`.lab/context/evidence-index.md\`
 
@@ -324,7 +331,15 @@ const ZH_SKILL_FILES = {
 - 必须把已批准的主指标、次级指标和必要终局证据明确写进 \`report.md\` 与受管的 \`main-tables.md\`。
 - 必须用白话解释选定的主指标和次级指标：每个指标在衡量什么、越高还是越低更好、它是主结果指标还是健康度/支持性指标。
 - 如果出现 coverage、completeness、confidence 或类似健康度指标，必须明确说明这类指标回答的是“实验是否跑稳、证据是否完整”，而不是主要科学效应本身。
+- 要把最关键的背景来源、方法/基线来源和指标来源直接写进报告，不要把它们藏在 \`.lab/context/*\` 里。
+- 如果 \`.lab/context/terminology-lock.md\` 里已经冻结了方法名和 contribution bullets，就必须把它们带进报告。
+- 方法概述必须用协作者能读懂的话说明：我们的方法大致怎么做、相对 closest prior work 或 strongest baseline 改了什么、这些 prior 方法各自做了什么，以及它们为什么在当前 claim 下仍然不够。
+- 只保留少量最关键的 prior work/baseline 锚点；每个锚点都要用一句话交代它做了什么和它的局限。
+- 在起草报告前，先检查 \`.lab/context/mission.md\` 和 \`.lab/context/eval-protocol.md\` 是否仍是模板空壳。
+- 如果 canonical context 还是空壳，要先根据 frozen result artifacts、data-decisions、evidence-index 和已批准上下文回填“最小可信版本”，再写报告。
+- 如果回填后仍缺少协作者可读所需的关键字段，就必须把输出降级成 \`artifact-anchored interim report\`，不能冒充最终协作者报告。
 - 如果报告依赖了对原始指标或原始实现的偏差，必须明确写出这个偏差。
+- workflow 工件状态、rerun id 或 LaTeX 骨架状态不能混进“已验证主结果”；这些内容必须单列到工件状态部分。
 - 如果 workflow language 是中文，\`report.md\` 和 \`<deliverables_root>/main-tables.md\` 也应使用中文，除非文件路径、代码标识符或字面指标名必须保持原样。
 - 解释优先保守，不要写成营销文案。
 - 要给 \`/lab:write\` 留下清晰 handoff，尤其是 section draft 可以直接引用的证据链接。
@@ -334,6 +349,7 @@ const ZH_SKILL_FILES = {
 - 开始前先简洁说明：campaign outcome、选定的主指标和次级指标、最强已支撑 claim、最大的报告风险。
 - 当该阶段由 \`/lab:auto\` 进入时，要主动给出用户可读的白话总结，不要等用户再追问“这些指标是什么意思”或“这些表怎么看”。
 - 把 \`report.md\` 当作给用户看的工件，而不是内部 dump。术语第一次出现时就解释；先讲结论，再讲术语。
+- 把 contribution bullets 当作协作者可读的最终主张摘要，而不是内部 TODO；每条都必须和当前证据边界对齐。
 - 如果某个未决前提会改变报告解释，一次只问一个问题。
 - 如果存在多种报告 framing，先给 2-3 个方案、trade-offs 和推荐项，优先最忠于证据的 framing。
 - 如果某种 framing 会实质影响后续论文 claim，要保留 approval gate。
@@ -691,6 +707,12 @@ const ZH_SKILL_FILES = {
   [path.join(".lab", ".managed", "templates", "final-report.md")]:
 `# 最终报告
 
+## 报告状态
+
+- 报告模式：collaborator-ready 或 artifact-anchored interim
+- canonical context 完整度：
+- 为什么当前只能用这个模式：
+
 ## 给用户看的总结
 
 - 一句话结论：
@@ -698,6 +720,33 @@ const ZH_SKILL_FILES = {
 - 还没有被证明的内容：
 - 当前最大报告风险：
 
+## 问题与背景
+
+- 这项研究在解决什么问题：
+- 为什么这个问题重要：
+- 当前报告到底覆盖了什么 setting 或 workflow：
+
+## 数据集场景说明
+
+- 数据集或 benchmark 1 代表什么真实场景：
+- 数据集或 benchmark 2 代表什么真实场景：
+- 数据集或 benchmark 3 代表什么真实场景：
+
+## 贡献总结
+
+- Contribution bullets：
+- 当前证据最强的贡献：
+- 仍需要更强证据的贡献：
+
+## 方法概述
+
+- 已批准的方法名：
+- 方法白话总结：
+- 相比 prior work 这套方法改变了什么：
+- 最相关的 prior work 或 baseline 锚点：
+- 这些 prior 方法各自做了什么：
+- 为什么这些 prior 方法在这里仍然不够：
+
 ## 选定指标
 
 - 主指标：
@@ -740,6 +789,11 @@ const ZH_SKILL_FILES = {
 - 最终表现摘要：
 - 主表覆盖情况：
 
+## 工件状态
+
+- 已就绪的交付物或工作流工件：
+- 这些工件状态为什么不是科学结论：
+
 ## 主要结果
 
 - 主要发现 1：
@@ -953,6 +1007,7 @@ const ZH_SKILL_FILES = {
 - 一句话问题：
 - 为什么重要：
 - 目标失败场景：
+- 回填来源：
 
 ## 成功标准
 
@@ -973,6 +1028,7 @@ const ZH_SKILL_FILES = {
 - 已批准方向：
 - 当前 owner 或会话：
 - 最近更新该 mission 的 stage：
+- 协作者可读状态：
 `,
   [path.join(".lab", "context", "state.md")]:
 `# 工作流状态
@@ -1966,6 +2022,7 @@ ZH_CONTENT[path.join(".lab", "context", "eval-protocol.md")] = `# 评估协议
 ## 指标释义
 
 - 指标释义：
+- 回填来源：
 - 背景来源：
 - 方法与基线来源论文：
 - 方法与基线实现来源：
@@ -2021,10 +2078,12 @@ ZH_CONTENT[path.join(".codex", "skills", "lab", "stages", "auto.md")] = `# \`/la
 
 ## 上下文写回
 
+- \`.lab/context/mission.md\`
 - \`.lab/context/state.md\`
 - \`.lab/context/decisions.md\`
 - \`.lab/context/data-decisions.md\`
 - \`.lab/context/evidence-index.md\`
+- \`.lab/context/eval-protocol.md\`
 - \`.lab/context/summary.md\`
 - \`.lab/context/session-brief.md\`
 - \`.lab/context/auto-status.md\`
@@ -2053,6 +2112,8 @@ ZH_CONTENT[path.join(".codex", "skills", "lab", "stages", "auto.md")] = `# \`/la
   - \`review\` 更新规范审查上下文
   - \`report\` 写出 \`<deliverables_root>/report.md\`
   - \`write\` 写出 \`<deliverables_root>/paper/\` 下的 LaTeX 产物
+- 如果即将进入 \`report\`，而 \`.lab/context/mission.md\` 或 \`.lab/context/eval-protocol.md\` 仍是模板空壳，就先根据冻结工件和已批准上下文回填最小可信版本。
+- 如果回填后仍缺少协作者可读所需的关键字段，就必须强制生成 \`artifact-anchored interim report\`，不能冒充最终协作者报告。
 - promotion 成功后，必须写回 \`data-decisions.md\`、\`decisions.md\`、\`state.md\` 和 \`session-brief.md\`。
 - 如果某个指标或对比 claim 在评估协议里没有带来源的定义，就不能拿它做 stop 或 promotion 判断。
 

package/package-assets/shared/lab/.managed/templates/final-report.md

@@ -1,5 +1,11 @@
 # Final Report
 
+## Report Status
+
+- Report mode: collaborator-ready or artifact-anchored interim
+- Canonical context readiness:
+- Why the current mode is appropriate:
+
 ## Reader Summary
 
 - One-sentence conclusion:
@@ -7,6 +13,33 @@
 - What is still unproven:
 - Biggest reporting risk:
 
+## Problem and Background
+
+- Research problem in plain language:
+- Why this problem matters:
+- What setting or workflow this report is actually about:
+
+## Dataset Scene Notes
+
+- Dataset or benchmark 1 and what real-world setting it represents:
+- Dataset or benchmark 2 and what real-world setting it represents:
+- Dataset or benchmark 3 and what real-world setting it represents:
+
+## Contribution Summary
+
+- Contribution bullets:
+- Strongest supported contribution:
+- Contributions that still need stronger evidence:
+
+## Method Overview
+
+- Approved method name:
+- Plain-language method summary:
+- What this method changes relative to prior work:
+- Most relevant prior work or baseline anchors:
+- What those prior methods do:
+- Why those prior methods are still insufficient here:
+
 ## Selected Metrics
 
 - Primary metrics:
@@ -56,6 +89,11 @@
 - Final performance summary:
 - Table coverage:
 
+## Artifact Status
+
+- Deliverables or workflow artifacts that are ready:
+- Artifact status notes that are not scientific findings:
+
 ## Main Results
 
 Summarize validated iteration outcomes.

package/package-assets/shared/lab/context/eval-protocol.md

@@ -17,6 +17,7 @@ Use this file to define the paper-facing evaluation objective, table plan, gates
 ## Metric Glossary
 
 - Metric glossary:
+- Hydration provenance:
 - Background sources:
 - Method and baseline source papers:
 - Method and baseline implementation source:

package/package-assets/shared/lab/context/mission.md

@@ -5,6 +5,7 @@
 - One-sentence problem:
 - Why it matters:
 - Target failure case:
+- Hydration provenance:
 
 ## Success Criteria
 
@@ -25,3 +26,4 @@
 - Approved direction:
 - Current owner or session:
 - Last stage that updated this mission:
+- Collaborator-ready status:

package/package-assets/shared/skills/lab/SKILL.md

@@ -46,6 +46,7 @@ Use this skill when the user invokes `/lab:*` or asks for the structured researc
 - Keep an explicit approval gate before `/lab:spec`.
 - Write idea artifacts with the template in `.lab/.managed/templates/idea.md`.
 - Update `.lab/context/mission.md`, `.lab/context/decisions.md`, and `.lab/context/open-questions.md` after convergence.
+- Do not leave `.lab/context/mission.md` as a template shell once the problem statement and approved direction are known.
 - Do not implement code in this stage.
 
 ### `/lab:data`
@@ -111,7 +112,8 @@ Use this skill when the user invokes `/lab:*` or asks for the structured researc
 - Normalize the result with `.lab/.managed/scripts/eval_report.py`.
 - Validate normalized output with `.lab/.managed/scripts/validate_results.py`.
 - Read `.lab/context/eval-protocol.md` before choosing the smallest run so the first experiment already targets the approved tables, metrics, and gates.
-- Update `.lab/context/state.md` and `.lab/context/evidence-index.md` after the run.
+- Update `.lab/context/state.md`, `.lab/context/evidence-index.md`, and `.lab/context/eval-protocol.md` after the run.
+- If the evaluation protocol is still skeletal, initialize the smallest trustworthy source-backed version before treating the run as the protocol anchor.
 
 ### `/lab:iterate`
 
@@ -131,7 +133,8 @@ Use this skill when the user invokes `/lab:*` or asks for the structured researc
 - Keep metric definitions, baseline behavior, and comparison implementations anchored to the source-backed evaluation protocol before changing thresholds, gates, or ladder transitions.
 - Switch to diagnostic mode if risk increases for two consecutive rounds.
 - Write round reports with `.lab/.managed/templates/iteration-report.md`.
-- Update `.lab/context/state.md`, `.lab/context/decisions.md`, `.lab/context/evidence-index.md`, and `.lab/context/open-questions.md` each round as needed.
+- Update `.lab/context/state.md`, `.lab/context/decisions.md`, `.lab/context/evidence-index.md`, `.lab/context/open-questions.md`, and `.lab/context/eval-protocol.md` each round as needed.
+- Keep `.lab/context/eval-protocol.md` synchronized with accepted ladder changes, benchmark scope, and source-backed implementation deviations.
 - Stop at threshold success or iteration cap, and record blockers plus next-best actions when the campaign ends without success.
 
 ### `/lab:review`
@@ -153,7 +156,9 @@ Use this skill when the user invokes `/lab:*` or asks for the structured researc
 - Aggregate them with `.lab/.managed/scripts/summarize_iterations.py`.
 - Write the final document with `.lab/.managed/templates/final-report.md` and the managed table summary with `.lab/.managed/templates/main-tables.md`.
 - Keep failed attempts and limitations visible.
-- Update `.lab/context/state.md` and `.lab/context/evidence-index.md` with report-level handoff notes.
+- Update `.lab/context/mission.md`, `.lab/context/eval-protocol.md`, `.lab/context/state.md`, and `.lab/context/evidence-index.md` with report-level handoff notes.
+- If canonical context is still skeletal, hydrate the smallest trustworthy version from frozen artifacts before finalizing the report.
+- If collaborator-critical fields remain missing after hydration, downgrade to an `artifact-anchored interim report` instead of presenting a final collaborator-ready report.
 
 ### `/lab:write`
 

package/package-assets/shared/skills/lab/stages/auto.md

@@ -24,6 +24,8 @@
 
 ## Context Write Set
 
+- `.lab/context/mission.md`
+- `.lab/context/eval-protocol.md`
 - `.lab/context/state.md`
 - `.lab/context/decisions.md`
 - `.lab/context/data-decisions.md`
@@ -63,6 +65,8 @@
   - keep the session alive while the current rung is running
   - write the current rung, watch target, and next rung to `.lab/context/auto-status.md`
 - Reuse the existing `/lab:run`, `/lab:iterate`, `/lab:review`, `/lab:report`, and optional `/lab:write` contracts instead of inventing a parallel workflow.
+- If the loop is about to reach `report` while `.lab/context/mission.md` or `.lab/context/eval-protocol.md` is still skeletal, hydrate the smallest trustworthy canonical version from frozen artifacts and approved context before drafting the report.
+- If hydration still leaves collaborator-critical fields blank, force `report` to emit an `artifact-anchored interim report` instead of a collaborator-ready final report.
 - Enforce stage contracts, not just exit codes:
   - `run` and `iterate` must change persistent outputs under `results_root`
   - `review` must update canonical review context

package/package-assets/shared/skills/lab/stages/idea.md

@@ -3,6 +3,7 @@
 ## Required Output
 
 - one-sentence problem statement
+- why the problem matters in plain language
 - failure case
 - idea classification
 - contribution category
@@ -18,6 +19,7 @@
 - critique before convergence
 - minimum viable experiment
 - explicit approval gate before `/lab:spec`
+- canonical mission context updated with the approved problem, importance, failure case, and direction
 
 ## Evidence Discipline
 
@@ -60,3 +62,4 @@
 - State why the target problem matters before talking about the method.
 - Compare against existing methods explicitly, not by vague novelty language.
 - The three meaningful points should each fit in one direct sentence.
+- Do not leave `.lab/context/mission.md` as an empty template after convergence; write the approved problem, why it matters, the current benchmark scope, and the approved direction back into canonical context.

package/package-assets/shared/skills/lab/stages/iterate.md

@@ -29,6 +29,7 @@ Declare and keep fixed:
 - `.lab/context/decisions.md`
 - `.lab/context/evidence-index.md`
 - `.lab/context/open-questions.md`
+- `.lab/context/eval-protocol.md`
 
 ## Per-Round Output
 
@@ -62,6 +63,7 @@ If the loop stops without success, record:
 - Do not accumulate long-lived results under `.lab/changes/<change-id>/runs`.
 - Do not change metric definitions, baseline semantics, or comparison implementations unless the approved evaluation protocol records both their sources and any deviations.
 - When you change ladders, sample sizes, or promotion gates, keep the resulting logic anchored to the source-backed evaluation protocol instead of ad-hoc chat reasoning.
+- Keep `.lab/context/eval-protocol.md` synchronized with the active benchmark scope, ladder gates, source-backed metric definitions, and any accepted implementation deviations instead of leaving it as a stale template.
 
 ## Interaction Contract
 

package/package-assets/shared/skills/lab/stages/report.md

@@ -2,7 +2,11 @@
 
 ## Required Output
 
+- report status: collaborator-ready or artifact-anchored interim
 - reader summary for the user
+- problem and background in plain language
+- dataset scene notes in plain language
+- contribution summary
 - method overview
 - selected metrics summary
 - plain-language metric guide
@@ -17,6 +21,7 @@
 - failed attempts
 - limitations
 - next steps
+- artifact status kept separate from validated findings
 
 ## Context Read Set
 
@@ -30,6 +35,8 @@
 
 ## Context Write Set
 
+- `.lab/context/mission.md`
+- `.lab/context/eval-protocol.md`
 - `.lab/context/state.md`
 - `.lab/context/evidence-index.md`
 
@@ -43,8 +50,15 @@
 - Explain the selected primary and secondary metrics in plain language for the user: what each metric measures, whether higher or lower is better, and whether it is a main result metric or only a health/support metric.
 - If coverage, completeness, confidence, or similar health metrics appear, explicitly say that they describe experimental reliability rather than the main scientific effect.
 - Pull the core background references, method or baseline references, and metric references out of the approved evaluation protocol instead of hiding them in `.lab/context/*`.
+- Pull the approved method name and contribution bullets out of `.lab/context/terminology-lock.md` when that framing context exists; do not silently drop them from the collaborator-facing report.
+- Explain the method overview in collaborator language: what the method roughly does, what changed relative to the closest prior work or strongest baseline, what those prior methods do, and why they remain insufficient for the approved claim.
+- When citing prior work or baselines in the method overview, include only the few anchor references a collaborator needs, and summarize their role and limitation in one short line each.
 - Report only the few references a collaborator needs to orient themselves quickly; do not turn `report.md` into a full bibliography dump.
 - If the report depends on a deviation from an original metric or implementation, state that deviation explicitly instead of smoothing it over.
+- Before drafting the report, inspect `.lab/context/mission.md` and `.lab/context/eval-protocol.md` for skeletal template fields.
+- If either canonical context file is still skeletal, hydrate the smallest trustworthy version from frozen result artifacts, dataset decisions, evidence-index, and prior approved context, and write that back before finalizing the report.
+- If collaborator-critical fields still remain missing after hydration, downgrade the output to an `artifact-anchored interim report` instead of presenting it as a final collaborator-ready report.
+- Do not mix workflow deliverable status, rerun ids, or manuscript skeleton status into validated scientific findings; keep those in a separate artifact-status section.
 - If `.lab/config/workflow.json` sets the workflow language to Chinese, write `report.md` and `<deliverables_root>/main-tables.md` in Chinese unless a file path, code identifier, or literal metric name must remain unchanged.
 - Prefer conservative interpretation over marketing language.
 - Leave a clear handoff path into `/lab:write` with evidence links that section drafts can cite.
@@ -54,6 +68,7 @@
 - Start with a concise summary of the campaign outcome, the selected primary and secondary metrics, the strongest supported claim, and the biggest reporting risk.
 - Proactively deliver a user-readable plain-language summary when the stage is reached from `/lab:auto`; do not wait for a separate follow-up request asking what the metrics or tables mean.
 - Treat `report.md` as a user-facing artifact rather than an internal dump. Prefer plain-language explanations before jargon, and explain each metric the first time it matters.
+- Treat contribution bullets as collaborator-facing claim summaries, not as internal TODOs; tie each one to the current evidence boundary.
 - If a missing assumption would change report interpretation, ask one clarifying question at a time.
 - If there are multiple defensible report framings, present 2-3 approaches with trade-offs and recommend the most evidence-faithful framing before writing.
 - Keep an approval gate when the reporting frame would materially affect what the paper later claims.

package/package-assets/shared/skills/lab/stages/run.md

@@ -6,6 +6,7 @@
 - run registry entry
 - normalized evaluation summary
 - validation result for the normalized summary
+- canonical evaluation context initialized or refined when the active protocol is still skeletal
 
 ## Context Read Set
 
@@ -19,6 +20,7 @@
 
 - `.lab/context/state.md`
 - `.lab/context/evidence-index.md`
+- `.lab/context/eval-protocol.md`
 
 ## Constraints
 
@@ -26,6 +28,7 @@
 - Fail fast on data, environment, or metric wiring problems.
 - Tie the run to the approved evaluation protocol, not just an ad-hoc chat goal.
 - Do not invent metric definitions, baseline behavior, or comparison implementations from memory; anchor them to the approved evaluation protocol and its recorded sources.
+- If `.lab/context/eval-protocol.md` is still skeletal, write the smallest trustworthy version of the current evaluation objective, metric set, ladder, and source-backed implementation notes before treating the run as the new protocol anchor.
 - Record the exact launch command and output location.
 - Write durable run outputs, logs, and checkpoints under `results_root`.
 - Write figures or plots under `figures_root`.

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "superlab",
-  "version": "0.1.18",
+  "version": "0.1.20",
   "description": "Strict /lab research workflow installer for Codex and Claude",
   "keywords": [
     "codex",