@tea-agent/loop-agent 0.19.0 → 0.20.0

package/CHANGELOG.md

@@ -2,6 +2,17 @@
 
 ## [Unreleased]
 
+## [0.20.0] - 2026-07-23
+
+### 新增
+
+- streamline implementation DAG (#61)
+- improve human-readable artifacts
+
+### 修复
+
+- relax traceability mapping
+
 ## [0.19.0] - 2026-07-23
 
 ### 重点更新
@@ -17,6 +28,7 @@
 
 ### 改进
 
+- `frontend-implementation` 保留独立 contract/scout，并把 Mock 策略与 implementation contract 合入 plan，把写前授权、验证/失败评估、复验和 review context 分别收敛为组合 shell 节点。standard/high-risk 固定 15 个顶层节点、small-risk 13 个；绿色路径执行 11 个节点、7 次 Pi，同时保留唯一 prewrite 写入授权、同 writeSet repair、真实 diff review 与 fail-closed closeout。
 - 优化操作与观测统一界面的导航与布局，合并确认与运行为「开始运行」，DAG 详情跳转路径更加准确
 
 ### 修复

package/dist/application/dag/generate-task-dag.js

@@ -105,6 +105,17 @@ function isVerificationShellTask(task) {
     const commands = resolveShellCommands(task.shell);
     return commands.some((command) => /(vitest|npm run (lint|typecheck|test)|check-repo\.sh|loop-agent-standard-verify)/.test(command));
 }
+function resolveReviewPacketShellCommands(task) {
+    const bundle = task.shell?.frontendVerificationBundle;
+    if (bundle) {
+        return [
+            ...bundle.mockCommands,
+            ...bundle.staticCommands,
+            ...bundle.behaviorCommands,
+        ];
+    }
+    return resolveShellCommands(task.shell);
+}
 async function buildReviewPacket(input) {
     const spec = parseDagSpec(JSON.parse(await readFile(input.dagPath, "utf-8")));
     const writers = collectWriterTasksForPacket(spec).map((task) => {
@@ -123,7 +134,7 @@ async function buildReviewPacket(input) {
         .filter(isVerificationShellTask)
         .map((task) => ({
         nodeId: task.id,
-        commands: resolveShellCommands(task.shell),
+        commands: resolveReviewPacketShellCommands(task),
     }));
     return {
         profileRouting: {

package/dist/executors/shell-executor.js

@@ -12,10 +12,12 @@ import { materializeFrontendTestResult, } from "../workflows/dag/frontend-test-r
 import { formatFrontendVerificationTraceStdout, runFrontendVerificationTraceGate, } from "../workflows/dag/frontend-verification-trace.js";
 import { formatFrontendWorktreeDiffStdout, runFrontendWorktreeDiffGate, } from "../workflows/dag/frontend-worktree-diff.js";
 import { formatFrontendFailureAssessStdout, formatFrontendRepairContractStdout, runFrontendFailureAssessGate, runFrontendRepairContractGate, } from "../workflows/dag/frontend-repair.js";
+import { formatFrontendPrewriteGateStdout, runFrontendPrewriteGate, } from "../workflows/dag/frontend-prewrite-gate.js";
+import { formatFrontendReviewContextStdout, runFrontendReviewContextGate, } from "../workflows/dag/frontend-review-context.js";
 import { formatTraceabilityGateStdout, materializeBackendTestCaseManifest, runBackendTestTraceabilityGate, } from "../workflows/dag/backend-test-case-manifest.js";
 import { materializeBackendTestExecutionContract } from "../workflows/dag/backend-test-execution-contract.js";
 import { materializeBackendTestResultFromRunDir, parseJunitXml } from "../workflows/dag/backend-test-result-contract.js";
-import { inspectBackendTestEnvironment, requiredBackendMarkdownCaseAcIds, renderBackendTestFacts, renderBackendTestHtml, redactBackendTestOutput, validateBackendMarkdownCases, validateBackendMarkdownTraceability, writeRunReport, } from "../workflows/dag/backend-test-markdown-workflow.js";
+import { collectBackendTestHumanCaseCatalog, inspectBackendTestEnvironment, requiredBackendMarkdownCaseAcIds, renderBackendTestFacts, renderBackendTestHtml, redactBackendTestOutput, validateBackendMarkdownCases, validateBackendMarkdownTraceability, writeRunReport, } from "../workflows/dag/backend-test-markdown-workflow.js";
 import { buildBackendTestCanonicalResultFromInitialShellSnippet, materializeBackendTestClassification, } from "../workflows/dag/backend-test-classification-contract.js";
 import { backendTestSemanticReviewSchema, materializeBackendTestSemanticReview, } from "../workflows/dag/backend-test-semantic-review-contract.js";
 import { pathsChangedDuringRun, readGitStatusPorcelain, snapshotGitStatusPorcelain, validateShellWriteGuard, } from "./shell-write-guard.js";
@@ -356,14 +358,16 @@ async function executeBackendTestPipeline(input, meta) {
             if (![0, 1].includes(pytestExitCode))
                 throw new Error(`pytest did not complete with a reportable exit code: ${pytestExitCode}`);
             const parsed = parseJunitXml(junitContent);
+            const cases = await collectBackendTestHumanCaseCatalog(input.cwd);
             const htmlContent = renderBackendTestHtml({
                 title: meta.spec.title,
                 parsed,
+                cases,
                 environmentSummary: await readFile(path.join(reportsDir, "backend-test-environment.md"), "utf8"),
                 traceabilitySummary: await readFile(path.join(reportsDir, "backend-test-traceability.md"), "utf8"),
             });
             const htmlPath = await writeRunReport(meta.runDir, "backend-test.html", htmlContent);
-            const facts = renderBackendTestFacts({ parsed, pytestExitCode, junitRelativePath: "reports/backend-test.junit.xml", htmlRelativePath: "reports/backend-test.html", junitContent, htmlContent });
+            const facts = renderBackendTestFacts({ parsed, cases, pytestExitCode, junitRelativePath: "reports/backend-test.junit.xml", htmlRelativePath: "reports/backend-test.html", junitContent, htmlContent });
             const factsPath = await writeRunReport(meta.runDir, "backend-test-facts.md", facts);
             const sanitizedOutputs = results.map((result) => redactBackendTestOutput(result.stdout));
             outputs.push(...sanitizedOutputs, `html=${htmlPath}`, `facts=${factsPath}`, facts);
@@ -601,8 +605,190 @@ async function executePipelineCommands(input, meta, overrideCommands) {
     }
     return results;
 }
+async function executeFrontendVerificationBundle(input, meta) {
+    const started = Date.now();
+    const shell = input.task.shell;
+    const bundle = shell.frontendVerificationBundle;
+    const cwd = resolveShellCwd(input.cwd, shell.cwd);
+    const results = [];
+    let beforeStatus;
+    try {
+        beforeStatus = await readGitStatusPorcelain(input.cwd);
+    }
+    catch {
+        beforeStatus = undefined;
+    }
+    const groups = [
+        { name: "mock", commands: bundle.mockCommands },
+        { name: "static", commands: bundle.staticCommands },
+        { name: "behavior", commands: bundle.behaviorCommands },
+    ];
+    for (const group of groups) {
+        for (const command of group.commands) {
+            const commandNumber = results.length + 1;
+            const result = await executeShellCommand({
+                command,
+                cwd,
+                timeoutMs: shell.timeoutMs ?? DEFAULT_SHELL_TIMEOUT_MS,
+                envAllowlist: shell.envAllowlist,
+                dagRunMeta: { runDir: meta.runDir, runId: meta.runId },
+                outputArtifacts: {
+                    stdoutPath: path.join(meta.runDir, input.task.id, "commands", `${commandNumber}.stdout.txt`),
+                    stderrPath: path.join(meta.runDir, input.task.id, "commands", `${commandNumber}.stderr.txt`),
+                },
+            });
+            results.push(result);
+            if (!result.ok)
+                break;
+        }
+    }
+    if (beforeStatus !== undefined) {
+        const guard = await runShellWriteGuard({
+            rootCwd: input.cwd,
+            task: input.task,
+            beforeStatus,
+        });
+        if (!guard.ok) {
+            return {
+                ok: false,
+                stdout: "",
+                stderr: `write guard failed: ${guard.violations.join(", ")}`,
+                failureCategory: "write-guard",
+                durationMs: Date.now() - started,
+            };
+        }
+    }
+    await writeDagNodeTextArtifact(meta.runDir, input.task.id, "result.summary.md", buildShellResultSummaryMarkdown({
+        nodeId: input.task.id,
+        runId: meta.runId,
+        rootCwd: input.cwd,
+        results,
+    }));
+    const commandResults = results.map((result) => ({
+        ok: result.ok,
+        exitCode: result.exitCode,
+        failureCategory: result.failureCategory,
+        command: result.command,
+    }));
+    const firstFailure = results.find((result) => !result.ok);
+    let traceError;
+    try {
+        await runFrontendVerificationTraceGate({
+            runDir: meta.runDir,
+            workspaceRoot: input.cwd,
+            evidence: {
+                static: {
+                    nodeId: input.task.id,
+                    commandLabels: bundle.staticEvidence.commandLabels,
+                },
+                behavior: {
+                    nodeId: input.task.id,
+                    commandLabels: bundle.behaviorEvidence.commandLabels,
+                },
+            },
+        });
+    }
+    catch (error) {
+        traceError = error instanceof Error ? error : new Error(String(error));
+    }
+    if (bundle.mode === "repair") {
+        if (firstFailure || traceError) {
+            return {
+                ok: false,
+                stdout: summarizeCommandResults(results).stdout,
+                stderr: firstFailure?.stderr || traceError?.message || "frontend reverify failed",
+                failureCategory: firstFailure?.failureCategory ?? "invalid-output",
+                durationMs: Date.now() - started,
+                ...{ commandResults },
+            };
+        }
+        return {
+            ok: true,
+            stdout: "Frontend reverify bundle: pass",
+            stderr: "",
+            failureCategory: "success",
+            durationMs: Date.now() - started,
+            ...{ commandResults },
+        };
+    }
+    const failureFacts = [];
+    if (firstFailure) {
+        failureFacts.push({
+            nodeId: input.task.id,
+            record: {
+                status: "FINISHED",
+                failureCategory: firstFailure.failureCategory,
+                stdout: summarizeCommandResults(results).stdout,
+                stderr: firstFailure.stderr,
+                commandResults,
+            },
+        });
+    }
+    if (traceError) {
+        failureFacts.push({
+            nodeId: "frontend-verification-trace-shell",
+            record: {
+                status: "FINISHED",
+                failureCategory: "invalid-output",
+                stderr: traceError.message,
+            },
+        });
+    }
+    try {
+        const assessment = await runFrontendFailureAssessGate({
+            runDir: meta.runDir,
+            failureFacts,
+        });
+        await runFrontendRepairContractGate({ runDir: meta.runDir });
+        return {
+            ok: true,
+            stdout: formatFrontendFailureAssessStdout(assessment),
+            stderr: "",
+            failureCategory: "success",
+            durationMs: Date.now() - started,
+            ...{ commandResults },
+        };
+    }
+    catch (error) {
+        return {
+            ok: false,
+            stdout: "",
+            stderr: error instanceof Error ? error.message : String(error),
+            failureCategory: "invalid-output",
+            durationMs: Date.now() - started,
+            ...{ commandResults },
+        };
+    }
+}
 export async function executeDagShellNode(input, meta) {
     const shell = input.task.shell;
+    if (shell?.frontendPrewriteGate) {
+        const started = Date.now();
+        try {
+            const result = await runFrontendPrewriteGate({
+                runDir: meta.runDir,
+                config: shell.frontendPrewriteGate,
+                sourceBinding: meta.spec.sourceBinding,
+            });
+            return { ok: true, stdout: formatFrontendPrewriteGateStdout(result), stderr: "", failureCategory: "success", durationMs: Date.now() - started };
+        }
+        catch (error) {
+            return { ok: false, stdout: "", stderr: error instanceof Error ? error.message : String(error), failureCategory: "invalid-output", durationMs: Date.now() - started };
+        }
+    }
+    if (shell?.frontendVerificationBundle) {
+        return executeFrontendVerificationBundle(input, meta);
+    }
+    if (shell?.frontendReviewContext) {
+        const started = Date.now();
+        try {
+            const result = await runFrontendReviewContextGate({ runDir: meta.runDir, workspaceRoot: input.cwd });
+            return { ok: true, stdout: formatFrontendReviewContextStdout(result), stderr: "", failureCategory: "success", durationMs: Date.now() - started };
+        }
+        catch (error) {
+            return { ok: false, stdout: "", stderr: error instanceof Error ? error.message : String(error), failureCategory: "invalid-output", durationMs: Date.now() - started };
+        }
+    }
     if (shell?.backendTestPipeline) {
         return executeBackendTestPipelineWithWriteGuard(input, meta);
     }

package/dist/workflows/dag/backend-test-markdown-workflow.js

@@ -5,13 +5,23 @@ const CASE_ID = /\bBE-[A-Z0-9]+(?:-[A-Z0-9]+)*-\d{3}\b/g;
 const AC_ID = /\bAC-[A-Z0-9]+(?:-[A-Z0-9]+)*\b/g;
 const SECRET = /(?:password|passwd|secret|token|api[_-]?key|private[_-]?key|authorization)\s*[:=]\s*\S+/i;
 const PLACEHOLDER = /\b(?:TODO|TBD|FIXME)\b|后续补充|暂不实现|接口正常|结果正确|按实际情况处理/i;
+const CASE_SECTION_ALIASES = {
+    acceptanceCriteria: ["Acceptance Criteria", "验收标准"],
+    sourceReferences: ["Source References", "需求依据"],
+    preconditions: ["Preconditions", "前置条件"],
+    testData: ["Test Data", "测试数据"],
+    steps: ["Steps", "操作步骤"],
+    expectedResults: ["Expected Results", "预期结果"],
+    automationNotes: ["Automation Notes", "自动化映射", "自动化说明"],
+    testPurpose: ["Test Purpose", "测试目的", "测试场景"],
+};
 const REQUIRED_CASE_SECTIONS = [
-    "Acceptance Criteria",
-    "Source References",
-    "Preconditions",
-    "Steps",
-    "Expected Results",
-    "Automation Notes",
+    CASE_SECTION_ALIASES.acceptanceCriteria,
+    CASE_SECTION_ALIASES.sourceReferences,
+    CASE_SECTION_ALIASES.preconditions,
+    CASE_SECTION_ALIASES.steps,
+    CASE_SECTION_ALIASES.expectedResults,
+    CASE_SECTION_ALIASES.automationNotes,
 ];
 async function exists(filePath) {
     try {
@@ -34,7 +44,7 @@ function unique(values) {
     return [...new Set(values)].sort((left, right) => left.localeCompare(right));
 }
 export function requiredBackendMarkdownCaseAcIds(criteria) {
-    const downstreamEvidence = /(?:pytest|junit|html|stdout|stderr|traceability|追溯|一一映射|一一追溯|执行并产生|执行结果|测试报告|失败分类|证据)/i;
+    const downstreamEvidence = /(?:pytest|junit|html|stdout|stderr|traceability|markdown|readme|human-readable|skip|xfail|吞(?:断言|异常)|生产代码|生产配置|配置写入|写入边界|中文(?:用例|文档|结构|展示)?|用例(?:索引|结构|文档)|测试目的|需求依据|操作步骤|预期结果|自动化映射|追溯|一一映射|一一追溯|执行并产生|执行结果|测试报告|失败分类|证据)/i;
     return unique([...criteria]
         .filter((criterion) => !downstreamEvidence.test(criterion))
         .flatMap((criterion) => criterion.match(/\bAC-[A-Z0-9]+(?:-[A-Z0-9]+)*\b/g) ?? []));
@@ -165,8 +175,9 @@ function splitCases(markdown) {
         body: markdown.slice(match.index, headings[index + 1]?.index ?? markdown.length),
     }));
 }
-function sectionBody(body, heading) {
-    const marker = new RegExp(`^###\\s+${heading}\\s*$`, "mi");
+function sectionBody(body, headings) {
+    const headingPattern = headings.map(escapeRegExp).join("|");
+    const marker = new RegExp(`^###\\s+(?:${headingPattern})\\s*$`, "mi");
     const match = marker.exec(body);
     if (!match)
         return "";
@@ -320,14 +331,15 @@ export async function validateBackendMarkdownCases(input) {
             if (previous)
                 throw new Error(`duplicate case id ${testCase.id}: ${previous} and ${path.relative(input.workspaceRoot, file)}`);
             seen.set(testCase.id, path.relative(input.workspaceRoot, file).replaceAll(path.sep, "/"));
-            for (const section of REQUIRED_CASE_SECTIONS) {
-                if (!new RegExp(`^###\\s+${section}\\s*$`, "mi").test(testCase.body)) {
-                    throw new Error(`${testCase.id} missing section: ${section}`);
+            for (const sectionAliases of REQUIRED_CASE_SECTIONS) {
+                const sectionPattern = sectionAliases.map(escapeRegExp).join("|");
+                if (!new RegExp(`^###\\s+(?:${sectionPattern})\\s*$`, "mi").test(testCase.body)) {
+                    throw new Error(`${testCase.id} missing section: ${sectionAliases.join(" or ")}`);
                 }
             }
             const acIds = unique(testCase.body.match(AC_ID) ?? []);
             acIds.forEach((id) => coveredAc.add(id));
-            const sourceSection = sectionBody(testCase.body, "Source References");
+            const sourceSection = sectionBody(testCase.body, CASE_SECTION_ALIASES.sourceReferences);
             const sourceRefs = extractSourceReferences({
                 sourceSection,
                 sourceBinding: input.sourceBinding,
@@ -348,8 +360,8 @@ export async function validateBackendMarkdownCases(input) {
                     throw new Error(`${testCase.id} references a missing source path: ${sourceRef}`);
                 }
             }
-            const steps = sectionBody(testCase.body, "Steps");
-            const expected = sectionBody(testCase.body, "Expected Results");
+            const steps = sectionBody(testCase.body, CASE_SECTION_ALIASES.steps);
+            const expected = sectionBody(testCase.body, CASE_SECTION_ALIASES.expectedResults);
             if (!hasNumberedListItem(steps))
                 throw new Error(`${testCase.id} has no numbered executable step`);
             if (!hasAssertableExpectedResult(expected))
@@ -381,18 +393,32 @@ function collectMarkdownCaseIds(markdown) {
 }
 function testFunctionRegion(input) {
     const functionLineStart = input.source.lastIndexOf("\n", input.functionIndex - 1) + 1;
-    const prefix = input.source.slice(0, functionLineStart);
-    const previousBlankLine = Math.max(prefix.lastIndexOf("\n\n"), prefix.lastIndexOf("\r\n\r\n"));
-    const decoratorCandidateStart = previousBlankLine >= 0
-        ? previousBlankLine + (prefix.startsWith("\r\n", previousBlankLine) ? 4 : 2)
-        : 0;
-    const decoratorCandidate = input.source.slice(decoratorCandidateStart, functionLineStart);
-    const regionStart = /^\s*@/m.test(decoratorCandidate)
-        ? decoratorCandidateStart
-        : functionLineStart;
-    const nextBoundary = /^(?:(?:async\s+)?def|class)\s+[A-Za-z_][A-Za-z0-9_]*\b/gm;
-    nextBoundary.lastIndex = input.functionHeaderEnd;
-    const regionEnd = nextBoundary.exec(input.source)?.index ?? input.source.length;
+    const functionLine = input.source.slice(functionLineStart, input.functionHeaderEnd);
+    const indent = functionLine.match(/^\s*/)?.[0] ?? "";
+    let regionStart = functionLineStart;
+    let cursor = functionLineStart;
+    while (cursor > 0) {
+        const previousEnd = cursor - 1;
+        const previousStart = input.source.lastIndexOf("\n", previousEnd - 1) + 1;
+        const previousLine = input.source.slice(previousStart, previousEnd).replace(/\r$/, "");
+        if (!previousLine.trim())
+            break;
+        if (new RegExp(`^${escapeRegExp(indent)}@`).test(previousLine)) {
+            regionStart = previousStart;
+            cursor = previousStart;
+            continue;
+        }
+        break;
+    }
+    const lines = input.source.slice(input.functionHeaderEnd).split(/\r?\n/);
+    let consumed = 0;
+    for (const line of lines) {
+        const boundary = line.match(/^(\s*)(?:(?:async\s+)?def|class)\s+[A-Za-z_][A-Za-z0-9_]*\b/);
+        if (boundary && boundary[1].length <= indent.length)
+            break;
+        consumed += line.length + 1;
+    }
+    const regionEnd = Math.min(input.source.length, input.functionHeaderEnd + consumed);
     return input.source
         .slice(regionStart, regionEnd)
         .split(/\r?\n/)
@@ -436,8 +462,8 @@ export async function validateBackendMarkdownTraceability(workspaceRoot) {
         if (unsafeSwallowedException) {
             throw new Error(`swallowed exception is forbidden in generated backend pytest: ${path.relative(workspaceRoot, file)} (${unsafeSwallowedException})`);
         }
-        for (const match of content.matchAll(/^(?:async\s+)?def\s+(test_[A-Za-z0-9_]+)\s*\([^)]*\)\s*(?:->\s*[^:\r\n]+)?\s*:/gm)) {
-            const symbol = match[1];
+        for (const match of content.matchAll(/^(\s*)(?:async\s+)?def\s+(test_[A-Za-z0-9_]+)\s*\([^)]*\)\s*(?:->\s*[^:\r\n]+)?\s*:/gm)) {
+            const symbol = match[2];
             const region = testFunctionRegion({
                 source: content,
                 functionIndex: match.index,
@@ -482,6 +508,82 @@ export async function validateBackendMarkdownTraceability(workspaceRoot) {
         "- skip/xfail findings: 0",
     ].join("\n") + "\n");
 }
+function cleanCaseTitle(rawHeading, id) {
+    return rawHeading
+        .replace(/^##\s+/, "")
+        .replace(new RegExp(`^${escapeRegExp(id)}\\s*(?:[｜|—–-]\\s*)?`), "")
+        .trim() || id;
+}
+export async function collectBackendTestHumanCaseCatalog(workspaceRoot) {
+    const files = await markdownFiles(workspaceRoot);
+    const catalog = new Map();
+    for (const file of files) {
+        if (path.basename(file).toLowerCase() === "readme.md")
+            continue;
+        const markdown = await readFile(file, "utf8");
+        for (const testCase of splitCases(markdown)) {
+            const heading = testCase.body.match(/^##\s+.*$/m)?.[0] ?? testCase.id;
+            const purpose = sectionBody(testCase.body, CASE_SECTION_ALIASES.testPurpose).trim();
+            const fallbackScenario = sectionBody(testCase.body, CASE_SECTION_ALIASES.expectedResults)
+                .replace(/^\s*(?:\d+[.)]|[-*+])\s+/gm, "")
+                .replace(/\s+/g, " ")
+                .trim();
+            catalog.set(testCase.id, {
+                id: testCase.id,
+                title: cleanCaseTitle(heading, testCase.id),
+                scenario: purpose.replace(/\s+/g, " ").trim() || fallbackScenario || "详见 Markdown 用例步骤与预期结果。",
+                markdownPath: path.relative(workspaceRoot, file).replaceAll(path.sep, "/"),
+                testFunctions: [],
+            });
+        }
+    }
+    const testcaseRoot = path.join(workspaceRoot, "testcase");
+    const pythonFiles = [];
+    async function walk(directory) {
+        for (const entry of await readdir(directory, { withFileTypes: true })) {
+            const absolute = path.join(directory, entry.name);
+            if (entry.isDirectory())
+                await walk(absolute);
+            else if (/^test_.*\.py$/i.test(entry.name))
+                pythonFiles.push(absolute);
+        }
+    }
+    if (await exists(testcaseRoot))
+        await walk(testcaseRoot);
+    for (const file of pythonFiles) {
+        const content = await readFile(file, "utf8");
+        for (const match of content.matchAll(/^(\s*)(?:async\s+)?def\s+(test_[A-Za-z0-9_]+)\s*\([^)]*\)\s*(?:->\s*[^:\r\n]+)?\s*:/gm)) {
+            const symbol = match[2];
+            const region = testFunctionRegion({ source: content, functionIndex: match.index, functionHeaderEnd: match.index + match[0].length });
+            const ids = new Set(region.match(CASE_ID) ?? []);
+            const fromSymbol = symbolCaseId(symbol);
+            if (fromSymbol)
+                ids.add(fromSymbol);
+            for (const id of ids) {
+                const item = catalog.get(id);
+                if (!item)
+                    continue;
+                item.scriptPath = path.relative(workspaceRoot, file).replaceAll(path.sep, "/");
+                if (!item.testFunctions.includes(symbol))
+                    item.testFunctions.push(symbol);
+            }
+        }
+    }
+    return [...catalog.values()].sort((left, right) => left.id.localeCompare(right.id));
+}
+function junitCaseId(name) {
+    return symbolCaseId(name) ?? name.match(CASE_ID)?.[0];
+}
+function humanStatus(status) {
+    return status === "passed" ? "通过" : status === "failure" ? "失败" : status === "error" ? "错误" : "跳过";
+}
+function formatDuration(durationMs) {
+    return durationMs === undefined ? "未记录" : `${(durationMs / 1000).toFixed(3)} 秒`;
+}
+function inferredScriptPath(classname) {
+    const moduleName = classname.split(".").filter((part) => part && !/^Test/.test(part)).join("/");
+    return `${moduleName || "unknown"}.py`;
+}
 export function redactBackendTestOutput(value) {
     return value
         .replace(/((?:password|passwd|secret|token|api[_-]?key|authorization)\s*[:=]\s*)\S+/gi, "$1[REDACTED]")
@@ -496,19 +598,28 @@ function escapeHtml(value) {
 }
 export function renderBackendTestHtml(input) {
     const passRate = input.parsed.tests > 0 ? input.parsed.passed / input.parsed.tests : 0;
-    const failures = input.parsed.failures.length
-        ? input.parsed.failures
-            .map((failure) => `<tr><td>${escapeHtml(failure.name)}</td><td>${escapeHtml(failure.kind)}</td><td><pre>${escapeHtml(failure.message)}</pre></td></tr>`)
-            .join("")
-        : '<tr><td colspan="3">No failures</td></tr>';
-    return `<!doctype html><html lang="zh-CN"><head><meta charset="utf-8"><title>${escapeHtml(input.title)}</title><style>body{font-family:system-ui,"Microsoft YaHei",sans-serif;margin:2rem;color:#172033}h1,h2{color:#17365d}.cards{display:grid;grid-template-columns:repeat(3,minmax(8rem,1fr));gap:.75rem}.card{padding:1rem;border:1px solid #ccd6e0;border-radius:.5rem;background:#f7f9fc}.bad{color:#b42318}table{width:100%;border-collapse:collapse}th,td{border:1px solid #ccd6e0;padding:.55rem;text-align:left;vertical-align:top}pre{white-space:pre-wrap;max-height:18rem;overflow:auto}</style></head><body><h1>${escapeHtml(input.title)}</h1><div class="cards"><div class="card">Total<br><strong>${input.parsed.tests}</strong></div><div class="card">Passed<br><strong>${input.parsed.passed}</strong></div><div class="card bad">Failed/Error<br><strong>${input.parsed.failed + input.parsed.errors}</strong></div><div class="card">Skipped<br><strong>${input.parsed.skipped}</strong></div><div class="card">Pass rate<br><strong>${(passRate * 100).toFixed(2)}%</strong></div><div class="card">Duration<br><strong>${input.parsed.durationMs ?? "unavailable"} ms</strong></div></div><h2>Environment</h2><pre>${escapeHtml(input.environmentSummary.slice(0, 4000))}</pre><h2>Traceability</h2><pre>${escapeHtml(input.traceabilitySummary.slice(0, 4000))}</pre><h2>Failures</h2><table><thead><tr><th>Test</th><th>Kind</th><th>Summary</th></tr></thead><tbody>${failures}</tbody></table></body></html>`;
+    const catalog = new Map((input.cases ?? []).map((item) => [item.id, item]));
+    const rows = input.parsed.cases.map((result) => {
+        const caseId = junitCaseId(result.name) ?? "未关联";
+        const item = catalog.get(caseId);
+        const scriptPath = item?.scriptPath ?? inferredScriptPath(result.classname);
+        const details = result.details || result.message || "";
+        const failureCell = result.status === "passed"
+            ? "—"
+            : `<strong>${escapeHtml(result.message || humanStatus(result.status))}</strong>${details ? `<details><summary>查看完整技术详情</summary><pre>${escapeHtml(details)}</pre></details>` : ""}`;
+        return `<tr class="result-${result.status}"><td><code>${escapeHtml(caseId)}</code></td><td><strong>${escapeHtml(item?.title ?? result.name)}</strong><small>${escapeHtml(item?.scenario ?? "未从 Markdown 用例提取场景说明。")}</small></td><td><code>${escapeHtml(scriptPath)}</code><small>${escapeHtml(result.name)}</small></td><td><span class="status ${result.status}">${humanStatus(result.status)}</span></td><td>${formatDuration(result.durationMs)}</td><td>${failureCell}</td></tr>`;
+    }).join("");
+    const conclusion = input.parsed.failed + input.parsed.errors === 0
+        ? `本轮测试通过：${input.parsed.passed} 条用例执行成功。`
+        : `本轮测试未通过：${input.parsed.failed + input.parsed.errors} 条用例失败或错误，请优先查看失败原因。`;
+    return `<!doctype html><html lang="zh-CN"><head><meta charset="utf-8"><meta name="viewport" content="width=device-width,initial-scale=1"><title>${escapeHtml(input.title)}</title><style>:root{color-scheme:light;--ink:#172033;--muted:#667085;--line:#d9e2ec;--panel:#f7f9fc;--brand:#17365d;--pass:#067647;--pass-bg:#ecfdf3;--fail:#b42318;--fail-bg:#fef3f2;--skip:#9a6700;--skip-bg:#fffaeb}*{box-sizing:border-box}body{margin:0;background:#eef2f6;color:var(--ink);font-family:system-ui,"Microsoft YaHei","PingFang SC",sans-serif;line-height:1.6}header,main{max-width:1440px;margin:auto}header{padding:2.5rem 2rem 1.25rem}main{padding:0 2rem 3rem}h1{margin:.25rem 0;color:var(--brand)}h2{margin-top:2rem;color:var(--brand)}.eyebrow{color:var(--muted);font-weight:700}.banner{padding:1rem 1.25rem;border-radius:.75rem;background:${input.parsed.failed + input.parsed.errors === 0 ? "var(--pass-bg)" : "var(--fail-bg)"};color:${input.parsed.failed + input.parsed.errors === 0 ? "var(--pass)" : "var(--fail)"};font-weight:700}.cards{display:grid;grid-template-columns:repeat(6,minmax(7rem,1fr));gap:.75rem;margin:1rem 0}.card{padding:1rem;border:1px solid var(--line);border-radius:.75rem;background:white}.card span,small{display:block;color:var(--muted)}.card strong{font-size:1.55rem}section{background:white;border:1px solid var(--line);border-radius:.9rem;padding:1.25rem;margin-top:1rem;box-shadow:0 8px 24px rgba(23,54,93,.04)}.table-wrap{overflow:auto}table{width:100%;border-collapse:collapse;min-width:1100px}th,td{border-bottom:1px solid var(--line);padding:.75rem;text-align:left;vertical-align:top}th{background:var(--panel);position:sticky;top:0}.status{display:inline-block;padding:.15rem .55rem;border-radius:999px;font-weight:700}.status.passed{color:var(--pass);background:var(--pass-bg)}.status.failure,.status.error{color:var(--fail);background:var(--fail-bg)}.status.skipped{color:var(--skip);background:var(--skip-bg)}code{white-space:nowrap}pre{white-space:pre-wrap;max-height:22rem;overflow:auto;background:#101828;color:#f2f4f7;padding:1rem;border-radius:.5rem}details{margin-top:.5rem}summary{cursor:pointer;color:var(--brand);font-weight:700}.secondary{display:grid;grid-template-columns:1fr 1fr;gap:1rem}.secondary pre{background:var(--panel);color:var(--ink)}@media(max-width:900px){.cards{grid-template-columns:repeat(2,1fr)}.secondary{grid-template-columns:1fr}header,main{padding-left:1rem;padding-right:1rem}}</style></head><body><header><div class="eyebrow">后端自动化测试报告</div><h1>${escapeHtml(input.title)}</h1><p>主要面向测试、研发和评审人员；执行事实来自同一次 pytest JUnit。</p></header><main><section><h2>测试结论</h2><div class="banner">${escapeHtml(conclusion)}</div><div class="cards"><div class="card"><span>用例总数</span><strong>${input.parsed.tests}</strong></div><div class="card"><span>通过</span><strong>${input.parsed.passed}</strong></div><div class="card"><span>失败</span><strong>${input.parsed.failed}</strong></div><div class="card"><span>错误</span><strong>${input.parsed.errors}</strong></div><div class="card"><span>跳过</span><strong>${input.parsed.skipped}</strong></div><div class="card"><span>通过率 / 总耗时</span><strong>${(passRate * 100).toFixed(2)}%</strong><small>${formatDuration(input.parsed.durationMs)}</small></div></div></section><section><h2>用例执行结果</h2><div class="table-wrap"><table><thead><tr><th>用例编号</th><th>用例名称及测试场景</th><th>自动化脚本</th><th>结果</th><th>耗时</th><th>失败原因</th></tr></thead><tbody>${rows || '<tr><td colspan="6">未发现可展示的 JUnit testcase。</td></tr>'}</tbody></table></div></section><section><h2>执行证据</h2><div class="secondary"><details><summary>环境信息</summary><pre>${escapeHtml(input.environmentSummary.slice(0, 8000))}</pre></details><details><summary>用例追溯信息</summary><pre>${escapeHtml(input.traceabilitySummary.slice(0, 8000))}</pre></details></div></section></main></body></html>`;
 }
 export function renderBackendTestFacts(input) {
     const passRate = input.parsed.tests > 0 ? input.parsed.passed / input.parsed.tests : 0;
     return ([
-        "# Backend Test Execution Facts",
+        "# 后端测试执行事实",
         "",
-        "## Execution",
+        "## 执行摘要",
         "",
         `- Status: ${input.pytestExitCode === 0 ? "passed" : "completed-with-failures"}`,
         `- Pytest exit code: ${input.pytestExitCode}`,
@@ -516,7 +627,7 @@ export function renderBackendTestFacts(input) {
         "- JUnit valid: yes",
         "- HTML valid: yes",
         "",
-        "## Counts",
+        "## 计数",
         "",
         `- Total: ${input.parsed.tests}`,
         `- Passed: ${input.parsed.passed}`,
@@ -525,20 +636,31 @@ export function renderBackendTestFacts(input) {
         `- Skipped: ${input.parsed.skipped}`,
         `- Pass rate: ${(passRate * 100).toFixed(2)}%`,
         "",
-        "## Evidence",
+        "## 逐条执行结果",
+        "",
+        "| 用例编号 | 用例名称 | 自动化脚本 | 测试函数 | 结果 | 耗时 | 失败原因 |",
+        "|---|---|---|---|---|---:|---|",
+        ...input.parsed.cases.map((result) => {
+            const caseId = junitCaseId(result.name) ?? "未关联";
+            const item = new Map((input.cases ?? []).map((entry) => [entry.id, entry])).get(caseId);
+            const script = item?.scriptPath ?? inferredScriptPath(result.classname);
+            return `| ${caseId} | ${item?.title ?? result.name} | \`${script}\` | \`${result.name}\` | ${humanStatus(result.status)} | ${formatDuration(result.durationMs)} | ${(result.message ?? "—").replaceAll("|", "\\|")} |`;
+        }),
+        "",
+        "## 证据",
         "",
         `- JUnit: ${input.junitRelativePath}`,
         `- JUnit SHA-256: ${createHash("sha256").update(input.junitContent).digest("hex")}`,
         `- HTML: ${input.htmlRelativePath}`,
         `- HTML SHA-256: ${createHash("sha256").update(input.htmlContent).digest("hex")}`,
         "",
-        "## Failed Tests",
+        "## 失败用例",
         "",
         ...(input.parsed.failures.length > 0
             ? input.parsed.failures.map((failure) => `- ${failure.name} [${failure.kind}]: ${failure.message}`)
             : ["- None"]),
         "",
-        "## Maturity Evidence",
+        "## 成熟度证据",
         "",
         "- Code coverage: unavailable unless a separate validated coverage artifact exists.",
         "- Stability: unavailable unless at least five independent runs are recorded.",