fix(tooling): 补全PR测试报告解析并修复并发测试

- 更新 gframework-pr-review 脚本以提取 CTRF 测试摘要和失败用例详情

- 修复 SettingsModelTests 在 NET9+ 下错误使用 Monitor 持锁的并发测试语义

- 同步 analyzer-warning-reduction 的 active todo 与 trace 真值

.agents/skills/gframework-pr-review/SKILL.md

@@ -62,6 +62,7 @@ The script should produce:
 - Pre-merge failed checks, if present
 - Latest MegaLinter status and any detailed issues posted by `github-actions[bot]`
 - Test summary, including failed-test signals when present
+- Detailed failed-test rows from GitHub Test Reporter / CTRF comments when the PR comment includes `Name` / `Failure Message` content
 - CLI support for writing full JSON to a file and printing only narrowed text sections to stdout
 - Parse warnings only when both the primary API source and the intended fallback signal are unavailable
 

.agents/skills/gframework-pr-review/scripts/fetch_current_pr_review.py

@@ -257,6 +257,11 @@ def strip_markdown_links(text: str) -> str:
     return re.sub(r"\[([^\]]+)\]\([^)]+\)", r"\1", text)
 
 
+def strip_markdown_images(text: str) -> str:
+    """Drop Markdown image syntax while keeping surrounding text readable."""
+    return re.sub(r"!\[[^\]]*\]\([^)]+\)", "", text)
+
+
 def extract_section(text: str, start_marker: str, end_markers: list[str]) -> str | None:
     """Extract text between a start marker and the earliest matching end marker."""
     start = text.find(start_marker)
@@ -486,43 +491,190 @@ def parse_megalinter_comment(comment_body: str) -> dict[str, Any]:
     return report
 
 
+def clean_markdown_table_cell(text: str) -> str:
+    """Normalize a Markdown table cell for structured parsing."""
+    cleaned = strip_markdown_images(strip_markdown_links(html.unescape(text)))
+    cleaned = cleaned.replace("\xa0", " ")
+    cleaned = cleaned.replace("**", "").replace("*", "").replace("`", "")
+    return collapse_whitespace(cleaned)
+
+
+def parse_int_from_text(text: str) -> int | None:
+    """Extract the first integer value from text."""
+    match = re.search(r"\d+", text)
+    return int(match.group(0)) if match else None
+
+
+def parse_duration_from_text(text: str) -> str:
+    """Extract a duration token from text when present."""
+    match = re.search(r"\d+(?:\.\d+)?(?:ms|s|m|h)", text)
+    if match is not None:
+        return match.group(0)
+
+    return collapse_whitespace(text)
+
+
+def parse_markdown_table(table_text: str) -> tuple[list[str], list[list[str]]]:
+    """Parse a Markdown table into header cells and row cells."""
+    lines = [line.strip() for line in table_text.splitlines() if line.strip().startswith("|")]
+    if len(lines) < 2:
+        return [], []
+
+    headers = [clean_markdown_table_cell(cell) for cell in lines[0].strip("|").split("|")]
+    rows: list[list[str]] = []
+    for line in lines[2:]:
+        cells = [clean_markdown_table_cell(cell) for cell in line.strip("|").split("|")]
+        if cells:
+            rows.append(cells)
+
+    return headers, rows
+
+
+def extract_markdown_table_after_heading(block: str, heading: str) -> tuple[list[str], list[list[str]]]:
+    """Extract the first Markdown table that appears after a heading."""
+    section = extract_section(block, heading, ["\n### ", "\n#### ", "\n<details>", "\n<table>", "\n<sub>"])
+    if section is None:
+        return [], []
+
+    table_match = re.search(r"(\|.*\|\n\|[-| :]+\|\n(?:\|.*\|\n?)*)", section, re.S)
+    if table_match is None:
+        return [], []
+
+    return parse_markdown_table(table_match.group(1))
+
+
+def normalize_stat_header(header: str) -> str:
+    """Normalize a human-readable stats header into a stable machine key."""
+    ascii_only = re.sub(r"[^A-Za-z]+", "", header).lower()
+    aliases = {
+        "tests": "tests",
+        "passed": "passed",
+        "failed": "failed",
+        "skipped": "skipped",
+        "pending": "pending",
+        "other": "other",
+        "flaky": "flaky",
+        "duration": "duration",
+    }
+    return aliases.get(ascii_only, ascii_only)
+
+
+def parse_stats_table(headers: list[str], rows: list[list[str]]) -> dict[str, Any]:
+    """Convert a parsed Markdown stats table into the report stats shape."""
+    if not headers or not rows:
+        return {}
+
+    first_row = rows[0]
+    stats: dict[str, Any] = {}
+    for header, value in zip(headers, first_row):
+        key = normalize_stat_header(header)
+        if not key:
+            continue
+
+        if key == "duration":
+            stats[key] = parse_duration_from_text(value)
+            continue
+
+        parsed_value = parse_int_from_text(value)
+        if parsed_value is not None:
+            stats[key] = parsed_value
+
+    return stats
+
+
+def normalize_failure_message(text: str) -> str:
+    """Normalize a failed-test message while preserving the meaningful lines."""
+    cleaned = html.unescape(text)
+    cleaned = re.sub(r"(?i)<br\s*/?>", "\n", cleaned)
+    cleaned = re.sub(r"</?(?:p|div|tbody|thead|tr|td|th|table)>", "\n", cleaned)
+    cleaned = re.sub(r"<[^>]+>", " ", cleaned)
+    lines = [collapse_whitespace(line) for line in cleaned.splitlines()]
+    meaningful_lines = [line for line in lines if line]
+    return "\n".join(meaningful_lines)
+
+
+def parse_failed_test_summary_list(block: str) -> list[str]:
+    """Parse the compact failed-tests summary list from CTRF details blocks."""
+    failed_tests_section = re.search(
+        r"<details><summary><strong>\s*Failed Tests.*?</summary>(?P<body>.*?)</details>",
+        block,
+        re.S,
+    )
+    if failed_tests_section is None:
+        return []
+
+    summary_body = strip_markdown_links(strip_markdown_images(html.unescape(failed_tests_section.group("body"))))
+    failed_tests: list[str] = []
+    for raw_line in summary_body.splitlines():
+        line = collapse_whitespace(raw_line)
+        if not line:
+            continue
+
+        if "arrow-right" in raw_line:
+            parts = [part.strip() for part in line.split("arrow-right") if part.strip()]
+            candidate = parts[-1] if parts else line
+        elif ">" in line:
+            candidate = line.split(">")[-1].strip()
+        else:
+            candidate = line
+
+        if candidate:
+            failed_tests.append(candidate)
+
+    return failed_tests
+
+
+def parse_failed_test_details(block: str) -> list[dict[str, str]]:
+    """Parse the detailed failed-test HTML table from GitHub Test Reporter comments."""
+    details: list[dict[str, str]] = []
+    table_section = re.search(
+        r"### ❌ \*\*Some tests failed!\*\*.*?<tbody>(?P<body>.*?)</tbody>",
+        block,
+        re.S,
+    )
+    if table_section is None:
+        return details
+
+    for name_cell, message_cell in re.findall(r"<tr>\s*<td>(.*?)</td>\s*<td>(.*?)</td>\s*</tr>", table_section.group("body"), re.S):
+        name = collapse_whitespace(strip_tags(html.unescape(name_cell))).lstrip("❌").strip()
+        failure_message = normalize_failure_message(message_cell)
+        if name:
+            details.append(
+                {
+                    "name": name,
+                    "failure_message": failure_message,
+                }
+            )
+
+    return details
+
+
 def parse_test_report(block: str) -> dict[str, Any]:
     """Parse a CTRF or GitHub test-reporter comment block."""
     report: dict[str, Any] = {
         "raw": block.strip(),
         "stats": {},
         "failed_tests": [],
+        "failed_test_details": [],
         "has_failed_tests": False,
     }
 
-    summary_row_match = re.search(
+    summary_headers, summary_rows = extract_markdown_table_after_heading(block, "### Summary")
-        r"\|\s*\*?\*?(\d+)\*?\*?\s*\|\s*\*?\*?(\d+)\*?\*?\s*\|\s*\*?\*?(\d+)\*?\*?\s*\|"
+    report["stats"] = parse_stats_table(summary_headers, summary_rows)
-        r"\s*\*?\*?(\d+)\*?\*?\s*\|\s*\*?\*?(\d+)\*?\*?\s*\|\s*\*?\*?(\d+)\*?\*?\s*\|\s*\*?\*?([^\|]+?)\*?\*?\s*\|",
-        block,
-    )
-    if summary_row_match is not None:
-        report["stats"] = {
-            "tests": int(summary_row_match.group(1)),
-            "passed": int(summary_row_match.group(2)),
-            "failed": int(summary_row_match.group(3)),
-            "skipped": int(summary_row_match.group(4)),
-            "other": int(summary_row_match.group(5)),
-            "flaky": int(summary_row_match.group(6)),
-            "duration": summary_row_match.group(7).strip(),
-        }
 
-    failed_tests_section = extract_section(
+    if not report["stats"]:
-        block,
+        build_headers, build_rows = extract_markdown_table_after_heading(block, "### build-and-test:")
-        "### Failed Tests",
+        report["stats"] = parse_stats_table(build_headers, build_rows)
-        ["### Slowest Tests", "### Insights", "<sub>", "[Github Test Reporter]"],
+
-    )
+    failed_test_details = parse_failed_test_details(block)
-    if failed_tests_section:
+    failed_test_names = parse_failed_test_summary_list(block)
-        lines = [line.strip("- ").strip() for line in failed_tests_section.splitlines()[1:] if line.strip()]
+    if not failed_test_names and failed_test_details:
-        report["failed_tests"] = lines
+        failed_test_names = [detail["name"] for detail in failed_test_details]
-        report["has_failed_tests"] = True
+
-    elif "No failed tests in this run." in block or "All tests passed!" in block:
+    report["failed_tests"] = failed_test_names
-        report["failed_tests"] = []
+    report["failed_test_details"] = failed_test_details
-        report["has_failed_tests"] = False
+    failed_count = int(report["stats"].get("failed", 0) or 0)
+    report["has_failed_tests"] = bool(failed_test_names or failed_test_details or failed_count > 0)
 
     return report
 
@@ -1103,8 +1255,17 @@ def format_text(
                 lines.append(f"- Report {index}: no structured test stats parsed")
 
             if report["has_failed_tests"]:
-                for failed_test in report["failed_tests"]:
+                failed_test_details = report.get("failed_test_details", [])
-                    lines.append(f"  Failed test: {truncate_text(failed_test, max_description_length)}")
+                if failed_test_details:
+                    for failed_test_detail in failed_test_details:
+                        lines.append(f"  Failed test: {truncate_text(failed_test_detail['name'], max_description_length)}")
+                        lines.append(
+                            "  Failure: "
+                            f"{truncate_text(failed_test_detail['failure_message'].replace(chr(10), ' | '), max_description_length)}"
+                        )
+                else:
+                    for failed_test in report["failed_tests"]:
+                        lines.append(f"  Failed test: {truncate_text(failed_test, max_description_length)}")
             else:
                 lines.append("  Failed tests: none reported")
 

GFramework.Game.Tests/Setting/SettingsModelTests.cs

@@ -137,12 +137,10 @@ public sealed class SettingsModelTests
         var migrationMapLock = lockField!.GetValue(model);
         Assert.That(migrationMapLock, Is.Not.Null);
 
-        Task initializeTask;
+        var tasks = WithSynchronizationLockHeld(migrationMapLock!, () =>
-        Task registerTask;
-        lock (migrationMapLock!)
         {
-            initializeTask = Task.Run(() => model.InitializeAsync());
+            var initializeTask = Task.Run(() => model.InitializeAsync());
-            registerTask = Task.Run(() => model.RegisterMigration(new TestLatestSettingsMigrationV2ToV3()));
+            var registerTask = Task.Run(() => model.RegisterMigration(new TestLatestSettingsMigrationV2ToV3()));
 
             Thread.Sleep(50);
 
@@ -151,7 +149,11 @@ public sealed class SettingsModelTests
                 Assert.That(initializeTask.IsCompleted, Is.False);
                 Assert.That(registerTask.IsCompleted, Is.False);
             });
-        }
+
+            return (initializeTask, registerTask);
+        });
+
+        var (initializeTask, registerTask) = tasks;
 
         await Task.WhenAll(initializeTask, registerTask);
 
@@ -171,6 +173,35 @@ public sealed class SettingsModelTests
         });
     }
 
+    /// <summary>
+    ///     以与被测代码相同的同步原语持有反射获取到的锁对象，避免在 .NET 9+ 上把 <see cref="System.Threading.Lock" />
+    ///     退化成 <see cref="Monitor" /> 语义，导致并发测试误判。
+    /// </summary>
+    /// <param name="syncRoot">通过反射读取到的私有锁字段。</param>
+    /// <typeparam name="TResult">持锁代码返回的结果类型。</typeparam>
+    /// <param name="action">持锁期间执行的断言与并发调度逻辑。</param>
+    /// <returns>持锁代码的返回值。</returns>
+    private static TResult WithSynchronizationLockHeld<TResult>(object syncRoot, Func<TResult> action)
+    {
+        ArgumentNullException.ThrowIfNull(syncRoot);
+        ArgumentNullException.ThrowIfNull(action);
+
+#if NET9_0_OR_GREATER
+        if (syncRoot is System.Threading.Lock typedLock)
+        {
+            using (typedLock.EnterScope())
+            {
+                return action();
+            }
+        }
+#endif
+
+        lock (syncRoot)
+        {
+            return action();
+        }
+    }
+
     private sealed class TestSettingsData : ISettingsData
     {
         public string Value { get; set; } = "default";

ai-plan/public/analyzer-warning-reduction/todos/analyzer-warning-reduction-tracking.md

@@ -6,12 +6,12 @@
 
 ## 当前恢复点
 
-- 恢复点编号：`ANALYZER-WARNING-REDUCTION-RP-079`
+- 恢复点编号：`ANALYZER-WARNING-REDUCTION-RP-080`
-- 当前阶段：`Phase 79`
+- 当前阶段：`Phase 80`
 - 当前焦点：
-  - `2026-04-27` 已按 PR `#295` 的 latest-head review 收口本地仍成立的问题：`AsyncExtensionsTests` 的 `ArgumentException.ParamName` 契约与 active `ai-plan` 文档过长问题
+  - `2026-04-27` 已补齐 `$gframework-pr-review` 对 GitHub Test Reporter / CTRF 测试摘要与失败用例详情的提取，当前文本输出可以直接显示失败测试名和 failure message 摘要
-  - 当前轮次已重新确认 `origin/main` 基线与 `HEAD` 同为 `617e0bf`；当前已提交 `HEAD` 的 stop metric 仍为 `30 / 50` files、`642` changed lines，本轮 PR review 同步尚未提交入该指标
+  - `SettingsModelTests.RegisterMigration_During_Cache_Rebuild_Should_Not_Leave_Stale_Type_Cache` 已按 `System.Threading.Lock` 的真实语义修正测试实现，并完成针对性 Release 测试验证
-  - 当前剩余 warning 主要集中在 `YamlConfigSchemaValidator*`、`YamlConfigLoader.cs` 与大批量 `MA0048` 文件名拆分；这些 slice 仍高于本轮 PR review 收口的低风险边界
+  - 当前剩余 warning 热点仍集中在 `YamlConfigSchemaValidator*`、`YamlConfigLoader.cs` 与大批量 `MA0048` 文件名拆分；这些 slice 仍高于本轮 PR review / test follow-up 的低风险边界
 
 ## 当前活跃事实
 
@@ -19,12 +19,10 @@
 - 当前 PR review 真值：
   - `python3 .agents/skills/gframework-pr-review/scripts/fetch_current_pr_review.py --json-output <current-pr-review-json>`
     - 最新结果：成功；当前分支对应 PR 为 `#295`
-    - latest-head review 在本轮修复前共有 `2` 条 open thread，分别指向 `GFramework.Core.Tests/Extensions/AsyncExtensionsTests.cs` 与 active `trace`
+    - 当前测试报告输出已能显示 `Summary` 统计、失败测试名称，以及 `Name / Failure Message` 表格中的关键信息
-  - 本轮已在本地收口上述两类问题，待提交后可重新执行 `$gframework-pr-review` 确认线程自动收口
+- 当前直接验证结果：
-- 提权后的直接验证当前确认为：
+  - `dotnet test GFramework.Game.Tests/GFramework.Game.Tests.csproj -c Release --filter "FullyQualifiedName~RegisterMigration_During_Cache_Rebuild_Should_Not_Leave_Stale_Type_Cache"`
-  - `dotnet build GFramework.Core.Tests/GFramework.Core.Tests.csproj -c Release`
+    - 最新结果：成功；`Failed: 0, Passed: 1, Skipped: 0, Total: 1`
-    - 最新结果：成功；`126 Warning(s)`、`0 Error(s)`
-    - 当前仍可见的既有 warning ID：`CS8766`、`CS8625`、`CS8602`、`CS8618`、`MA0048`、`MA0002`、`MA0008`
 - 当前分支 stop-condition 指标：
   - `git diff --name-only refs/remotes/origin/main...HEAD | wc -l`
     - 最新结果：`30`
@@ -33,6 +31,7 @@
 - 当前批次摘要：
   - 三轮低风险 warning 清理已在此前验证中将仓库根 warning 从 `639` 降到 `397`
   - 当前批次的已完成 slice 明细已迁移到归档，active todo 仅保留恢复真值
+  - 本轮新增内容为 PR review 工具链补强与单个失败测试修正，不扩展 warning reduction 的热点清理边界
 - 当前建议保留到下一波次的候选：
   - `GFramework.Game/Config/YamlConfigLoader.cs` 的 `MA0158`（单点可修，但文件本身同时承载其他高耦合 warning）
   - 测试项目中的 `MA0048` 文件名拆分波次（会显著增加 changed-file 数）
@@ -43,8 +42,8 @@
   - 缓解措施：本轮先避开该热点，只清理低风险且 ownership 清晰的文件集合。
 - `MA0158` 迁移涉及 `net8.0` / `net9.0` / `net10.0` 多目标兼容。
   - 缓解措施：复用 `StoreSelection.cs` 已存在的 `#if NET9_0_OR_GREATER` 专用锁模式，不在 `net8.0` 引入不兼容 API。
-- 当前 PR open thread 的自动收口仍依赖新提交进入远端 PR head。
+- 当前 PR open thread 与 CI 失败信号仍依赖新提交进入远端 PR head 才能复核。
-  - 缓解措施：本轮提交后重新执行 `$gframework-pr-review`，仅以最新 head review 为准。
+  - 缓解措施：本轮提交后重新执行 `$gframework-pr-review`，同时确认 review thread 与 failed test signal 是否一起收口。
 
 ## 活跃文档
 
@@ -69,6 +68,6 @@
 
 ## 下一步建议
 
-1. 提交本轮 PR review 修复与 `ai-plan` 精简同步，再重跑 `$gframework-pr-review` 验证 PR `#295` 的 open thread 是否随新 head 收口。
+1. 提交本轮 `$gframework-pr-review` 解析增强、`SettingsModelTests` 修复与 `ai-plan` 同步。
-2. 若后续继续推进 warning reduction，建议另开下一波次，优先明确是否接受 `YamlConfigLoader.cs` 热点触碰，或是否要专门做测试项目 `MA0048` 拆分波次。
+2. 推送后重新执行 `$gframework-pr-review`，确认 PR `#295` 的 failed test detail 与 open thread 是否已更新为新 head 真值。
-3. 默认在当前恢复点停下，因为继续推进已不再符合本轮“PR review 收口 + 少文件同步”的边界。
+3. 若后续继续推进 warning reduction，建议另开下一波次处理 `YamlConfigLoader.cs` 热点或测试项目 `MA0048` 拆分波次。

ai-plan/public/analyzer-warning-reduction/traces/analyzer-warning-reduction-trace.md

@@ -1,37 +1,38 @@
 # Analyzer Warning Reduction 追踪
 
-## 2026-04-27 — RP-079
+## 2026-04-27 — RP-080
 
-### 阶段：按 PR `#295` latest-head review 收口当前仍成立的 open thread
+### 阶段：补强 `$gframework-pr-review` 的测试报告提取并修复 `SettingsModelTests` 失败用例
 
 - 触发背景：
-  - 重新执行 `$gframework-pr-review` 后，确认当前分支对应的最新公开 PR 为 `#295`，而不是旧 trace 中记录的 `#291`
+  - 用户补充了 GitHub Test Reporter / CTRF 的完整 PR 评论，指出现有 `$gframework-pr-review` 输出虽然能显示 `failed=1`，但没有抓到失败测试名称与详细报错
-  - latest-head review 仍有 `2` 条 open thread，分别指向 `GFramework.Core.Tests/Extensions/AsyncExtensionsTests.cs` 的参数名契约问题，以及 active trace 过长问题
+  - PR 评论中的真实失败用例为 `RegisterMigration_During_Cache_Rebuild_Should_Not_Leave_Stale_Type_Cache`，报错集中在测试通过反射持锁后两个任务提前完成
 - 主线程实施：
-  - 修正 `ThrowShouldNotRetry` 中 `ArgumentException` 的 `ParamName` 传递逻辑，并补充断言锁定 `nameof(taskFactory)` 契约
+  - 扩展 `.agents/skills/gframework-pr-review/scripts/fetch_current_pr_review.py` 的 `parse_test_report`，增加 CTRF `Summary` 统计、紧凑 failed-tests 列表、以及 `Name` / `Failure Message` HTML 表格的提取
-  - 将 active trace 精简为单一恢复入口，并把 `RP073` 到 `RP078` 的详细过程迁入归档
+  - 更新 `gframework-pr-review` skill 文档的输出预期，明确脚本现在应提取 GitHub Test Reporter / CTRF 的失败详情
-  - 同步压缩 active todo，只保留当前恢复点真值与归档指针
+  - 修正 `GFramework.Game.Tests/Setting/SettingsModelTests.cs`，让反射取得的 `_migrationMapLock` 在 `NET9_0_OR_GREATER` 下通过 `System.Threading.Lock.EnterScope()` 持锁，而不是退化成 `Monitor` 语义
 - 验证里程碑：
-  - `python3 .agents/skills/gframework-pr-review/scripts/fetch_current_pr_review.py --json-output <current-pr-review-json>`
+  - `python3 -c "... parse_test_report(...)"`（基于 `/tmp/current-pr-review.json` 的现有原始评论）
-    - 结果：成功；确认 PR `#295` latest-head review 在本轮修复前共有 `2` 条 open thread
+    - 结果：成功；已能解析 `tests=2156`、`passed=2155`、`failed=1`、`duration=35.3s`，并抓到 `RegisterMigration_During_Cache_Rebuild_Should_Not_Leave_Stale_Type_Cache` 的 failure message
-  - `dotnet build GFramework.Core.Tests/GFramework.Core.Tests.csproj -c Release`
+  - `python3 .agents/skills/gframework-pr-review/scripts/fetch_current_pr_review.py --section tests --json-output /tmp/current-pr-review-v2.json`
-    - 结果：成功；`126 Warning(s)`、`0 Error(s)`
+    - 结果：成功；文本输出已直接显示失败测试名与报错摘要
-    - 当前可见的既有 warning ID：`CS8766`、`CS8625`、`CS8602`、`CS8618`、`MA0048`、`MA0002`、`MA0008`
+  - `dotnet test GFramework.Game.Tests/GFramework.Game.Tests.csproj -c Release --filter "FullyQualifiedName~RegisterMigration_During_Cache_Rebuild_Should_Not_Leave_Stale_Type_Cache"`
+    - 结果：成功；`Failed: 0, Passed: 1, Skipped: 0, Total: 1`
 - 当前结论：
-  - 当前仍成立的 latest-head review 问题已经在本地收口，下一步应以新提交刷新 PR head 再复查
+  - `$gframework-pr-review` 现在已经能把 PR 测试评论里的关键失败用例信息直接提出来，不再只停留在 `failed=1`
-  - active `ai-plan` 入口现已回到“当前真值 + 归档指针”的体量，符合仓库对恢复文档的要求
+  - `SettingsModelTests` 当前失败是测试锁语义与生产代码不一致导致的误判，已在本地复现并修正通过
 
 ## 活跃风险
 
-- PR review thread 是否自动关闭仍取决于新提交是否进入远端 PR head。
+- PR 上的 latest-head review thread 与测试报告仍需要等新提交进入远端后再复核。
-  - 缓解措施：提交后重新执行 `$gframework-pr-review`，只以新的 latest-head 结果为准。
+  - 缓解措施：提交并推送后重新执行 `$gframework-pr-review`，只以新的 latest-head 和 test report 为准。
 - `YamlConfigSchemaValidator*`、`YamlConfigLoader.cs` 与 `MA0048` 拆分仍是下一波次的高耦合候选。
-  - 缓解措施：保持本轮边界只处理 PR review 收口，不顺手扩展 warning reduction 范围。
+  - 缓解措施：保持本轮边界只处理 PR review 工具链与失败测试，不顺手扩展 warning reduction 范围。
 
 ## 下一步
 
-1. 完成本轮验证并提交。
+1. 完成本轮提交。
-2. 重新执行 `$gframework-pr-review`，确认 PR `#295` latest-head 是否还有 unresolved thread。
+2. 推送后重新执行 `$gframework-pr-review`，确认 PR `#295` 的 failed test detail 与 unresolved thread 是否已刷新。
 
 ## 历史归档指针