feat: enhance findings and exclusions handling with repo state logging

Reviewed-on: #105

.amazonq/rules/triage-findings.md

@@ -7,7 +7,7 @@ When the task is to triage review findings, follow this workflow:
 3. Sort by severity: `critical` -> `warning` -> `info`.
 4. Renumber from 1 after sorting.
 5. Fix real issues with the smallest safe change.
-6. Add false positives to `.gitea/ai-review/exclusions.json`.
+6. Add false positives to `.gitea/ai-review/exclusions.json`, preserving the original wording, language, and semantics as much as possible.
 7. Add or update tests when behavior changes.
 8. Re-check the issue after each fix.
 

.claude/skills/triage-findings/SKILL.md

@@ -18,7 +18,7 @@ description: Triage findings, fix real issues, and exclude false positives.
    - info
 3. Renumber from 1.
 4. Fix real issues.
-5. Put false positives into `.gitea/ai-review/exclusions.json`.
+5. Put false positives into `.gitea/ai-review/exclusions.json`, preserving the original wording, language, and semantics as much as possible.
 6. Add tests when behavior changes.
 
 ## Output Rules
@@ -26,3 +26,4 @@ description: Triage findings, fix real issues, and exclude false positives.
 - Keep the final list short.
 - Keep numbering contiguous.
 - Preserve file path, location, and fix.
+- When writing exclusions, prefer the original issue text over paraphrased rewrites.

.codex/skills/triage-findings/SKILL.md

@@ -21,7 +21,7 @@ It is also used when some findings are false positives and should be moved into
 4. Renumber the sorted list from 1 upward.
 5. Rewrite each finding concisely so the final list reads cleanly and consistently.
 6. If a finding is a false positive, do not keep it in the final list.
-7. Add false positives to the exclusions list using the existing schema in the repo or task context.
+7. Add false positives to the exclusions list using the existing schema in the repo or task context, and preserve the original finding wording as much as possible, including language and semantics.
 
 ## Resolution Flow
 
@@ -41,4 +41,5 @@ After the list is merged and ordered, resolve the remaining findings one by one.
 - Keep numbering contiguous after filtering and merging.
 - Preserve useful details like file path, location, and suggested fix.
 - Keep exclusions entries minimal and consistent with the project schema.
+- When writing exclusions, prefer the original issue text and language; only paraphrase if needed to fit the schema.
 - If the source already provides a severity or title, keep it unless it conflicts with the final ordering.

.gemini/skills/triage-findings/SKILL.md

@@ -18,7 +18,7 @@ description: Triage findings, fix real issues, and exclude false positives.
    - info
 3. Renumber from 1.
 4. Fix real issues.
-5. Put false positives into `.gitea/ai-review/exclusions.json`.
+5. Put false positives into `.gitea/ai-review/exclusions.json`, preserving the original wording, language, and semantics as much as possible.
 6. Add tests when behavior changes.
 
 ## Output Rules
@@ -26,3 +26,4 @@ description: Triage findings, fix real issues, and exclude false positives.
 - Keep the final list short.
 - Keep numbering contiguous.
 - Preserve file path, location, and fix.
+- When writing exclusions, prefer the original issue text over paraphrased rewrites.

.github/copilot-instructions.md

@@ -7,7 +7,7 @@ Use the triage-finding workflow for review issue lists:
 3. Sort by severity: `critical` -> `warning` -> `info`.
 4. Renumber from 1.
 5. Fix real issues with the smallest safe change.
-6. Put false positives into `.gitea/ai-review/exclusions.json`.
+6. Put false positives into `.gitea/ai-review/exclusions.json`, preserving the original wording, language, and semantics as much as possible.
 7. Add or update tests when behavior changes.
 8. Re-check after each fix.
 

.github/skills/triage-findings/SKILL.md

@@ -7,7 +7,7 @@ Use the triage-finding workflow for review issue lists:
 3. Sort by severity: `critical` -> `warning` -> `info`.
 4. Renumber from 1.
 5. Fix real issues with the smallest safe change.
-6. Put false positives into `.gitea/ai-review/exclusions.json`.
+6. Put false positives into `.gitea/ai-review/exclusions.json`, preserving the original wording, language, and semantics as much as possible.
 7. Add or update tests when behavior changes.
 8. Re-check after each fix.
 

CLAUDE.md

@@ -7,7 +7,7 @@ When the task is to triage review findings, follow this workflow:
 3. Sort by severity: `critical` -> `warning` -> `info`.
 4. Renumber from 1 after sorting.
 5. Fix real issues with the smallest safe change.
-6. Add false positives to `.gitea/ai-review/exclusions.json`.
+6. Add false positives to `.gitea/ai-review/exclusions.json`, preserving the original wording, language, and semantics as much as possible.
 7. Add or update tests when behavior changes.
 8. Re-check the issue after each fix.
 

Dockerfile

@@ -11,6 +11,7 @@ COPY app/package.json /action/app/
 RUN cd /action/app && npm install
 
 COPY .amazonq/ /action/.amazonq/
+COPY .codex/ /action/.codex/
 COPY .claude/ /action/.claude/
 COPY .gemini/ /action/.gemini/
 COPY .github/ /action/.github/

GEMINI.md

@@ -7,7 +7,7 @@ Use the triage-finding workflow for review issue lists:
 3. Sort by severity: `critical` -> `warning` -> `info`.
 4. Renumber from 1.
 5. Fix real issues with the smallest safe change.
-6. Put false positives into `.gitea/ai-review/exclusions.json`.
+6. Put false positives into `.gitea/ai-review/exclusions.json`, preserving the original wording, language, and semantics as much as possible.
 7. Add or update tests when behavior changes.
 8. Re-check after each fix.
 

README.md

@@ -6,8 +6,8 @@
 
 1. 服務名稱、模型名稱、角色資訊(個性、符合個性的英文名稱、工作內容)，Comment 到 Push Request
 2. 每個角色個別分析 Git Diff 的內容產生新問題表格(問題等級、角色名稱、問題位置或行數、修改建議)
-3. 讀取所有未解決的舊問題(問題檔案 `.gitea/ai-review/findings.json` 存在於使用此 Action 的專案固定位置)加上新問題後，去除重複產生本次 Push Request 的問題表格(PR問題表格)覆蓋問題檔案
+3. 讀取來源分支中的所有未解決舊問題(問題檔案 `.gitea/ai-review/findings.json`)加上新問題後，去除重複產生本次 Push Request 的問題表格(PR問題表格)覆蓋問題檔案
-4. 讀取排除問題檔案(`.gitea/ai-review/exclusions.json` 存在於使用此 Action 的專案固定位置)，用來過濾PR問題表格中不需要處理的問題
+4. 讀取來源分支中的排除問題檔案(`.gitea/ai-review/exclusions.json`)，用來過濾PR問題表格中不需要處理的問題
 5. 從PR問題表格中取出所有舊問題，依照等級排序後 Comment 到 Push Request
 6. 從PR問題表格中取出所有新問題，排除嚴重等級的問題後 Comment 到 Push Request
 7. 從PR問題表格中取出所有新問題，將每個嚴重等級的問題 Comment 到 Push Request
@@ -23,8 +23,9 @@
 5. 將提示詞放到 ./app/prompts 內供程式讀取
 6. API Key 支援逗號分隔傳入多個，隨機順序各嘗試一次，全部失敗則 exit 1
 7. 讀取 Git Diff 時排除 `.gitea/`、`.amazonq/`、`.claude/`、`.codex/`、`.gemini/`、`.github/` 資料夾，以及 `CLAUDE.md`、`GEMINI.md`、`TODO.md`、`README.md`，避免 AI 分析 workflow 設定、skill 入口與文件等非業務程式碼
-8. 階段七驗證 `findings.json` 與 `exclusions.json` 是否為合法 JSON 格式，格式錯誤時先嘗試透過 AI 修正內容，再重新驗證；修正後仍不合法才 exit 1；之後才檢查檔案是否存在，不存在則建立並寫入 `[]`
+8. 階段七驗證來源分支中的 `findings.json` 與 `exclusions.json` 是否為合法 JSON 格式，格式錯誤時先嘗試透過 AI 修正內容，再重新驗證；修正後仍不合法才 exit 1；之後才檢查檔案是否存在，不存在則建立並寫入 `[]`
 9. 傳給 AI 的 findings 只保留必要欄位（level、role、location、suggestion），排除 `is_new` 等內部欄位；system prompt 精簡為指令核心；exclusions hint 只傳 location 與 suggestion，減少 token 用量
+10. 執行時會額外記錄來源分支狀態、`findings.json` / `exclusions.json` 的檔案路徑、大小、mtime 與 raw/normalized 筆數，方便追查讀檔與分支內容不一致的問題
 
 # 使用說明
 
@@ -227,4 +228,4 @@ Amazon Q：直接輸入 `triage-findings 問題原始檔(文字或截圖)`
 
 ### 版本包含
 
-提交時一併包含 `triage-findings` skill 與各平台入口檔；已存在檔案一律覆蓋，同步到最新內容；若 workspace 沒有某個同步檔，記憶區會保留原檔，不做刪除。
+提交時一併包含 `triage-findings` skill 與各平台入口檔；已存在檔案一律覆蓋，同步到最新內容；若 workspace 沒有某個同步檔，記憶區會保留原檔，不做刪除。`findings.json` 與 `exclusions.json` 都從使用此 action 的存取庫來源分支讀取，而不是從 action 本地 workspace 讀取。寫入 `.gitea/ai-review/exclusions.json` 時，盡量保留原始問題文字的語言與語意，避免過度改寫。未來若新增任何 skill 或新增其他平台的 skill 入口，必須同時把對應檔案複製進 Docker image，並把同步清單更新到會使用此 action 的目標專案，避免 action 與目標專案內容脫節。

TODO.md

@@ -40,13 +40,13 @@
 ## 階段八：記憶區 commit/push 與錯誤處理
 - 目標：記憶區能成功 commit/push，且一併包含 `triage-findings` skill 與各平台入口檔；skill 檔案已存在時一律以來源覆蓋，workspace 沒有的同步檔則保留記憶區既有內容，不做刪除；錯誤時有明確 log，流程結束有總結訊息。
 - 驗收：log 有「persisted findings」、「commit=...」、「push=...」等訊息，且能看出 skill 相關檔案已一併提交並被來源覆蓋；當 workspace 缺少某個同步檔時，記憶區中的對應檔案不會被刪除；錯誤時有「Runner failed: ...」等明確錯誤說明。
-- 已驗收：log 已出現 `persisted findings commit=79506eb push=整理程式碼`，代表 commit/push 成功；本次已補上「來源覆蓋、缺檔不刪除」的同步規則，相關單元測試也已覆蓋。
+- 已驗收：log 已出現 `persisted findings commit=b867eaa push=feat/解決問題`，代表 commit/push 成功；本次已補上「來源覆蓋、缺檔不刪除」的同步規則，相關單元測試也已覆蓋。
 
 ## 階段九：阻擋嚴重問題 PR（第 8 點）
 - 目標：如果 PR 問題表格中有嚴重（critical）問題，workflow 需直接 exit 1，不讓流程成功。
 - 驗收：log 中能看到「critical 問題存在，workflow 結束（exit 1）」等明確訊息，且 workflow 狀態為失敗。
-- 部分驗收：這次 log 顯示 `✅ 無嚴重問題`，因此只驗到正常放行路徑；`exit 1` 的阻擋分支仍需另一次含 critical 的 PR log 驗證。
+- 已驗收：這次 log 已明確出現 `❌ 發現 2 個嚴重問題，workflow 結束（exit 1）`，且 job 以失敗結束，證明阻擋分支確實生效。
-- 可驗收紀錄情境：只要 `Step8` 出現 `發現 X 個嚴重問題，workflow 結束（exit 1）`，且 job 以失敗結束，就能驗收這一項；如果該次 PR 的 `filtered` 清單含 `critical`，就應該會看到這段 log。
+- 補充紀錄：`Step8` 的退出訊息屬於預期行為，不代表 Step7 commit/push 失敗。
 
 ## 階段十：API Key 輪替
 - 目標：所有平台的 API Key 支援逗號分隔傳入多個，隨機順序各嘗試一次，單一 Key 失敗時自動換下一個，全部失敗則 exit 1。

app/comments.js

@@ -16,13 +16,19 @@ function buildTable(findings) {
 }
 
 /**
- * 寫入 findings.json 到 workspace
+ * 寫入 findings.json。
+ * 預設寫到 workspace；若提供 mirrorDir，則同步寫入另一份供 repo commit 使用。
  */
-export function saveFindings(workspace, findings) {
+export function saveFindings(workspace, findings, mirrorDir = null) {
-  const fullPath = path.join(workspace, FINDINGS_PATH);
+  const targets = [workspace];
-  fs.mkdirSync(path.dirname(fullPath), { recursive: true });
+  if (mirrorDir && mirrorDir !== workspace) targets.push(mirrorDir);
-  fs.writeFileSync(fullPath, JSON.stringify(findings, null, 2) + '\n', 'utf8');
+
-  console.log(`  ✅ findings 寫入: ${fullPath} (${findings.length} 筆)`);
+  for (const targetDir of targets) {
+    const fullPath = path.join(targetDir, FINDINGS_PATH);
+    fs.mkdirSync(path.dirname(fullPath), { recursive: true });
+    fs.writeFileSync(fullPath, JSON.stringify(findings, null, 2) + '\n', 'utf8');
+    console.log(`  ✅ findings 寫入: ${fullPath} (${findings.length} 筆)`);
+  }
 }
 
 /**

app/comments.test.js

@@ -0,0 +1,75 @@
+import { describe, it, afterEach } from 'node:test';
+import assert from 'node:assert/strict';
+import fs from 'node:fs';
+import os from 'node:os';
+import path from 'node:path';
+import { saveFindings } from './comments.js';
+import { FINDINGS_PATH } from './config.js';
+
+describe('saveFindings', () => {
+  const tempDirs = [];
+  const makeTempDir = prefix => {
+    const dir = fs.mkdtempSync(path.join(os.tmpdir(), prefix));
+    tempDirs.push(dir);
+    return dir;
+  };
+
+  it('writes findings to workspace and mirror dirs when provided', () => {
+    const workspace = makeTempDir('findings-ws-');
+    const mirrorDir = makeTempDir('findings-mirror-');
+    const findings = [{ level: 'warning', role: 'Leo', location: 'file.js:1', suggestion: 'test' }];
+
+    saveFindings(workspace, findings, mirrorDir);
+
+    const workspaceText = fs.readFileSync(path.join(workspace, FINDINGS_PATH), 'utf8');
+    const mirrorText = fs.readFileSync(path.join(mirrorDir, FINDINGS_PATH), 'utf8');
+    assert.equal(workspaceText, JSON.stringify(findings, null, 2) + '\n');
+    assert.equal(mirrorText, JSON.stringify(findings, null, 2) + '\n');
+  });
+
+  it('writes only to workspace when mirrorDir is omitted', () => {
+    const workspace = makeTempDir('findings-ws-');
+    const findings = [{ level: 'info', role: 'Maya', location: 'file.js:2', suggestion: 'note' }];
+
+    saveFindings(workspace, findings);
+
+    const workspaceText = fs.readFileSync(path.join(workspace, FINDINGS_PATH), 'utf8');
+    assert.equal(workspaceText, JSON.stringify(findings, null, 2) + '\n');
+  });
+
+  it('does not duplicate writes when mirrorDir matches workspace', () => {
+    const workspace = makeTempDir('findings-same-');
+    const findings = [];
+    const writeCalls = [];
+    const originalWriteFileSync = fs.writeFileSync;
+
+    fs.writeFileSync = (...args) => {
+      writeCalls.push(args[0]);
+      return originalWriteFileSync(...args);
+    };
+
+    try {
+      saveFindings(workspace, findings, workspace);
+    } finally {
+      fs.writeFileSync = originalWriteFileSync;
+    }
+
+    assert.equal(writeCalls.length, 1);
+    assert.equal(writeCalls[0], path.join(workspace, FINDINGS_PATH));
+  });
+
+  it('writes an empty JSON array when findings is empty', () => {
+    const workspace = makeTempDir('findings-empty-');
+
+    saveFindings(workspace, []);
+
+    const workspaceText = fs.readFileSync(path.join(workspace, FINDINGS_PATH), 'utf8');
+    assert.equal(workspaceText, '[]\n');
+  });
+
+  afterEach(() => {
+    while (tempDirs.length > 0) {
+      fs.rmSync(tempDirs.pop(), { recursive: true, force: true });
+    }
+  });
+});

app/findings.js

@@ -34,11 +34,30 @@ function readJSONArray(fullPath, label) {
   }
 }
 
+function normalizeExclusions(data) {
+  if (Array.isArray(data)) return data;
+  if (data && Array.isArray(data.excluded_findings)) return data.excluded_findings;
+  return [];
+}
+
+function formatFileTime(mtimeMs) {
+  if (!Number.isFinite(mtimeMs)) return 'unknown';
+  return new Date(mtimeMs).toISOString();
+}
+
 /**
- * 讀取舊 findings（從 workspace 的 FINDINGS_PATH）
+ * 讀取舊 findings（從來源分支的 cloned repoDir 中的 FINDINGS_PATH）
  */
 export function loadOldFindings(workspace) {
-  const old = readJSONArray(path.join(workspace, FINDINGS_PATH), '舊 findings ').map(f => ({ ...f, is_new: false }));
+  const fullPath = path.join(workspace, FINDINGS_PATH);
+  const old = readJSONArray(fullPath, '舊 findings ').map(f => ({ ...f, is_new: false }));
+  if (fs.existsSync(fullPath)) {
+    const stat = fs.statSync(fullPath);
+    console.log(`  讀取舊 findings 檔案: ${fullPath}`);
+    console.log(`  舊 findings 檔案資訊: bytes=${stat.size} mtime=${formatFileTime(stat.mtimeMs)} path=${path.relative(workspace, fullPath) || fullPath}`);
+  } else {
+    console.log(`  舊 findings 檔案不存在: ${fullPath}`);
+  }
   console.log(`  讀取舊 findings: ${old.length} 筆`);
   return old;
 }
@@ -104,11 +123,39 @@ export async function deduplicateWithAI(findings) {
 }
 
 /**
- * 讀取排除問題檔案（從 workspace 的 EXCLUSIONS_PATH）
+ * 讀取排除問題檔案（從來源分支的 cloned repoDir 中的 EXCLUSIONS_PATH）
  */
-export function loadExclusions(workspace) {
+export function loadExclusions(workspace, repoState = null) {
-  const exclusions = readJSONArray(path.join(workspace, EXCLUSIONS_PATH), '排除問題');
+  const fullPath = path.join(workspace, EXCLUSIONS_PATH);
-  console.log(`  讀取排除問題: ${exclusions.length} 筆`);
+  if (!fs.existsSync(fullPath)) {
+    console.log(`  排除問題檔案不存在，視為空: ${fullPath}`);
+    if (repoState) {
+      const branch = repoState.branch || 'detached';
+      const shortSha = repoState.shortSha || repoState.headSha || 'unknown';
+      console.log(`  來源分支狀態: branch=${branch} commit=${shortSha} commit_time=${repoState.commitTime || 'unknown'}`);
+    }
+    console.log('  讀取排除問題: raw=0 normalized=0 筆');
+    return [];
+  }
+
+  let exclusions = [];
+  let rawCount = 0;
+  try {
+    const stat = fs.statSync(fullPath);
+    const data = JSON.parse(fs.readFileSync(fullPath, 'utf8'));
+    rawCount = Array.isArray(data) ? data.length : Array.isArray(data?.excluded_findings) ? data.excluded_findings.length : 0;
+    exclusions = normalizeExclusions(data);
+    const branch = repoState?.branch || 'detached';
+    const shortSha = repoState?.shortSha || repoState?.headSha || 'unknown';
+    const commitTime = repoState?.commitTime || 'unknown';
+    console.log(`  讀取排除問題檔案: ${fullPath}`);
+    console.log(`  來源分支狀態: branch=${branch} commit=${shortSha} commit_time=${commitTime}`);
+    console.log(`  檔案資訊: bytes=${stat.size} mtime=${formatFileTime(stat.mtimeMs)} raw=${rawCount} normalized=${exclusions.length} path=${path.relative(workspace, fullPath) || fullPath}`);
+  } catch (e) {
+    console.log(`  ⚠️  讀取排除問題失敗: ${e.message}，視為空: ${fullPath}`);
+    exclusions = [];
+  }
+  console.log(`  讀取排除問題: raw=${rawCount} normalized=${exclusions.length} 筆`);
   return exclusions;
 }
 

app/findings.test.js

@@ -0,0 +1,98 @@
+import { describe, it, beforeEach, afterEach } from 'node:test';
+import assert from 'node:assert/strict';
+import fs from 'node:fs';
+import os from 'node:os';
+import path from 'node:path';
+import { loadOldFindings, loadExclusions, applyExclusions } from './findings.js';
+import { EXCLUSIONS_PATH, FINDINGS_PATH } from './config.js';
+
+describe('findings exclusions', () => {
+  let workspace;
+  let logs;
+  let originalLog;
+
+  beforeEach(() => {
+    workspace = fs.mkdtempSync(path.join(os.tmpdir(), 'findings-test-'));
+    logs = [];
+    originalLog = console.log;
+    console.log = (...args) => {
+      logs.push(args.join(' '));
+    };
+  });
+
+  afterEach(() => {
+    console.log = originalLog;
+    fs.rmSync(workspace, { recursive: true, force: true });
+  });
+
+  it('loads excluded_findings wrapper format', () => {
+    const fullPath = path.join(workspace, EXCLUSIONS_PATH);
+    fs.mkdirSync(path.dirname(fullPath), { recursive: true });
+    fs.writeFileSync(fullPath, JSON.stringify({
+      excluded_findings: [
+        { location: 'entrypoint.sh:180', title: 'fetch_package_versions jq overhead' },
+      ],
+    }, null, 2));
+
+    const exclusions = loadExclusions(workspace);
+
+    assert.equal(exclusions.length, 1);
+    assert.equal(exclusions[0].location, 'entrypoint.sh:180');
+    assert.equal(exclusions[0].title, 'fetch_package_versions jq overhead');
+  });
+
+  it('applies exclusions loaded from wrapper format', () => {
+    const findings = [
+      { location: 'entrypoint.sh:180', role: 'Maya', suggestion: 'keep' },
+      { location: 'README.md:12', role: 'Maya', suggestion: 'keep' },
+    ];
+    const exclusions = [
+      { location: 'entrypoint.sh:180', title: 'fetch_package_versions jq overhead' },
+    ];
+
+    const filtered = applyExclusions(findings, exclusions);
+
+    assert.equal(filtered.length, 1);
+    assert.equal(filtered[0].location, 'README.md:12');
+  });
+
+  it('logs exclusions file metadata and repo state when loading exclusions', () => {
+    const fullPath = path.join(workspace, EXCLUSIONS_PATH);
+    fs.mkdirSync(path.dirname(fullPath), { recursive: true });
+    fs.writeFileSync(fullPath, JSON.stringify([
+      { location: 'entrypoint.sh:180', suggestion: 'ignore' },
+      { location: 'README.md:12', suggestion: 'ignore' },
+    ], null, 2));
+
+    const repoState = {
+      branch: 'feat/test',
+      shortSha: 'abc1234',
+      commitTime: '2026-05-15T09:29:49.817Z',
+      repoDir: path.join(workspace, 'repo'),
+    };
+
+    const exclusions = loadExclusions(workspace, repoState);
+
+    assert.equal(exclusions.length, 2);
+    assert.ok(logs.some(line => line.includes(`讀取排除問題檔案: ${fullPath}`)));
+    assert.ok(logs.some(line => line.includes('來源分支狀態: branch=feat/test commit=abc1234')));
+    assert.ok(logs.some(line => line.includes('raw=2 normalized=2')));
+    assert.ok(logs.some(line => line.includes(`path=${path.relative(workspace, fullPath)}`)));
+  });
+
+  it('logs findings file metadata when loading old findings', () => {
+    const fullPath = path.join(workspace, FINDINGS_PATH);
+    fs.mkdirSync(path.dirname(fullPath), { recursive: true });
+    fs.writeFileSync(fullPath, JSON.stringify([
+      { level: 'info', role: 'Maya', location: 'README.md:12', suggestion: 'keep' },
+    ], null, 2));
+
+    const findings = loadOldFindings(workspace);
+
+    assert.equal(findings.length, 1);
+    assert.equal(findings[0].is_new, false);
+    assert.ok(logs.some(line => line.includes(`讀取舊 findings 檔案: ${fullPath}`)));
+    assert.ok(logs.some(line => line.includes('舊 findings 檔案資訊: bytes=')));
+    assert.ok(logs.some(line => line.includes(`path=${path.relative(workspace, fullPath)}`)));
+  });
+});

app/git.js

@@ -5,10 +5,12 @@ import { fileURLToPath } from 'url';
 import { GITEA_SERVER_URL, GITEA_REPOSITORY, GITEA_TOKEN, PR_HEAD_BRANCH, FINDINGS_PATH } from './config.js';
 
 const ACTION_ROOT = path.resolve(path.dirname(fileURLToPath(import.meta.url)), '..');
+const GENERATED_SYNC_PATHS = [FINDINGS_PATH, '.gitea/ai-review/exclusions.json'];
 const remoteUrl = `${GITEA_SERVER_URL.replace(/\/$/, '')}/${GITEA_REPOSITORY}.git`;
 export const SYNC_PATHS = [
-  FINDINGS_PATH,
   '.amazonq/rules/triage-findings.md',
+  '.codex/skills/triage-findings/SKILL.md',
+  '.codex/skills/triage-findings/agents/openai.yaml',
   '.claude/skills/triage-findings/SKILL.md',
   '.gemini/skills/triage-findings/SKILL.md',
   '.github/copilot-instructions.md',
@@ -39,6 +41,23 @@ function withAskpass(workspace, fn) {
   }
 }
 
+function readGitOutput(run, args, cwd, env) {
+  try {
+    return run(args, cwd, env);
+  } catch {
+    return '';
+  }
+}
+
+export function getRepoState(repoDir, _spawnSync = spawnSync) {
+  const run = makeRunner(_spawnSync);
+  const headSha = readGitOutput(run, ['rev-parse', 'HEAD'], repoDir);
+  const shortSha = readGitOutput(run, ['rev-parse', '--short', 'HEAD'], repoDir);
+  const branch = readGitOutput(run, ['branch', '--show-current'], repoDir);
+  const commitTime = readGitOutput(run, ['show', '-s', '--format=%cI', 'HEAD'], repoDir);
+  return { repoDir, branch, headSha, shortSha, commitTime };
+}
+
 /**
  * Clone PR head branch to workspace/repo (idempotent)
  */
@@ -66,6 +85,10 @@ export async function commitAndPush(workspace, repoDir, _spawnSync = spawnSync,
     await withAskpass(workspace, async credEnv => {
       run(['config', 'user.email', 'ai-review[bot]@gitea'], repoDir);
       run(['config', 'user.name', 'AI Review Bot'], repoDir);
+      if (PR_HEAD_BRANCH) {
+        run(['fetch', 'origin', PR_HEAD_BRANCH], repoDir, credEnv);
+        run(['reset', '--hard', `origin/${PR_HEAD_BRANCH}`], repoDir);
+      }
 
       const existingSyncPaths = [];
 
@@ -84,6 +107,16 @@ export async function commitAndPush(workspace, repoDir, _spawnSync = spawnSync,
       if (existingSyncPaths.length > 0) {
         run(['add', ...existingSyncPaths], repoDir);
       }
+      const generatedSyncPaths = GENERATED_SYNC_PATHS.filter(relPath => fs.existsSync(path.join(workspace, relPath)));
+      if (generatedSyncPaths.length > 0) {
+        for (const relPath of generatedSyncPaths) {
+          const src = path.join(workspace, relPath);
+          const dest = path.join(repoDir, relPath);
+          fs.mkdirSync(path.dirname(dest), { recursive: true });
+          fs.copyFileSync(src, dest);
+        }
+        run(['add', ...generatedSyncPaths], repoDir);
+      }
 
       const status = run(['status', '--porcelain'], repoDir);
       if (!status) {
@@ -93,8 +126,12 @@ export async function commitAndPush(workspace, repoDir, _spawnSync = spawnSync,
 
       const out = run(['commit', '-m', 'chore: update ai-review findings [skip ci]'], repoDir);
       const commitHash = out.match(/\[.+ ([a-f0-9]+)\]/)?.[1] || 'unknown';
-      run(['push', remoteUrl, PR_HEAD_BRANCH], repoDir, credEnv);
+      try {
-      console.log(`  ✅ persisted findings  commit=${commitHash}  push=${PR_HEAD_BRANCH}`);
+        run(['push', remoteUrl, PR_HEAD_BRANCH], repoDir, credEnv);
+        console.log(`  ✅ persisted findings  commit=${commitHash}  push=${PR_HEAD_BRANCH}`);
+      } catch (pushErr) {
+        console.log(`  ⚠️  Step7 commit 成功但 push 失敗: commit=${commitHash} push=${PR_HEAD_BRANCH} error=${pushErr.message}`);
+      }
     });
   } catch (e) {
     console.log(`  ⚠️  Runner failed: commit/push 失敗: ${e.message}`);

app/git.test.js

@@ -94,18 +94,31 @@ describe('commitAndPush', () => {
   });
 
   it('adds skill and entry files together with findings', async () => {
+    const repoDir = path.join(workspace, 'repo');
+    fs.mkdirSync(path.join(workspace, '.gitea/ai-review'), { recursive: true });
+    fs.writeFileSync(path.join(workspace, '.gitea/ai-review/findings.json'), '[]\n');
+    fs.writeFileSync(path.join(workspace, '.gitea/ai-review/exclusions.json'), '[]\n');
+    fs.mkdirSync(path.join(repoDir, '.gitea/ai-review'), { recursive: true });
+    fs.writeFileSync(path.join(repoDir, '.gitea/ai-review/findings.json'), '[]\n');
+    fs.writeFileSync(path.join(repoDir, '.gitea/ai-review/exclusions.json'), '[]\n');
     const spawn = makeSpawn();
-    await commitAndPush(workspace, path.join(workspace, 'repo'), spawn, sourceRoot);
+    await commitAndPush(workspace, repoDir, spawn, sourceRoot);
-    const addCall = spawn.calls.find(c => c.args[0] === 'add');
+    const addCalls = spawn.calls.filter(c => c.args[0] === 'add');
-    assert.ok(addCall, 'expected git add to run');
+    const skillAddCall = addCalls.find(c => c.args.includes('.github/skills/triage-findings/SKILL.md'));
-    assert.ok(addCall.args.includes('.github/skills/triage-findings/SKILL.md'));
+    const generatedAddCall = addCalls.find(c => c.args.includes('.gitea/ai-review/exclusions.json'));
-    assert.ok(addCall.args.includes('.claude/skills/triage-findings/SKILL.md'));
+    assert.ok(skillAddCall, 'expected git add for synced skill files');
-    assert.ok(addCall.args.includes('.gemini/skills/triage-findings/SKILL.md'));
+    assert.ok(generatedAddCall, 'expected git add for generated review files');
-    assert.ok(addCall.args.includes('.github/copilot-instructions.md'));
+    assert.ok(skillAddCall.args.includes('.codex/skills/triage-findings/SKILL.md'));
-    assert.ok(addCall.args.includes('.amazonq/rules/triage-findings.md'));
+    assert.ok(skillAddCall.args.includes('.codex/skills/triage-findings/agents/openai.yaml'));
-    assert.ok(addCall.args.includes('CLAUDE.md'));
+    assert.ok(skillAddCall.args.includes('.claude/skills/triage-findings/SKILL.md'));
-    assert.ok(addCall.args.includes('GEMINI.md'));
+    assert.ok(skillAddCall.args.includes('.gemini/skills/triage-findings/SKILL.md'));
-    assert.ok(!addCall.args.includes('README.md'));
+    assert.ok(skillAddCall.args.includes('.github/copilot-instructions.md'));
+    assert.ok(skillAddCall.args.includes('.amazonq/rules/triage-findings.md'));
+    assert.ok(skillAddCall.args.includes('CLAUDE.md'));
+    assert.ok(skillAddCall.args.includes('GEMINI.md'));
+    assert.ok(!skillAddCall.args.includes('README.md'));
+    assert.ok(generatedAddCall.args.includes('.gitea/ai-review/findings.json'));
+    assert.ok(generatedAddCall.args.includes('.gitea/ai-review/exclusions.json'));
   });
 
   it('keeps repo copies when the source sync file is missing', async () => {
@@ -137,6 +150,32 @@ describe('commitAndPush', () => {
     const failSpawn = () => ({ status: 1, stdout: '', stderr: 'fatal: error', error: null });
     await assert.doesNotReject(() => commitAndPush(workspace, path.join(workspace, 'repo'), failSpawn, sourceRoot));
   });
+
+  it('logs push failures separately from commit failures', async () => {
+    const repoDir = path.join(workspace, 'repo');
+    fs.mkdirSync(path.join(workspace, '.gitea/ai-review'), { recursive: true });
+    fs.writeFileSync(path.join(workspace, '.gitea/ai-review/findings.json'), '[]\n');
+    fs.writeFileSync(path.join(workspace, '.gitea/ai-review/exclusions.json'), '[]\n');
+    fs.mkdirSync(path.join(repoDir, '.gitea/ai-review'), { recursive: true });
+    fs.writeFileSync(path.join(repoDir, '.gitea/ai-review/findings.json'), '[]\n');
+    fs.writeFileSync(path.join(repoDir, '.gitea/ai-review/exclusions.json'), '[]\n');
+
+    const spawn = makeSpawn({
+      push: () => ({ status: 1, stdout: '', stderr: 'remote: error: pre-receive hook declined', error: null }),
+    });
+    const logs = [];
+    const originalLog = console.log;
+    console.log = (...args) => { logs.push(args.join(' ')); };
+
+    try {
+      await commitAndPush(workspace, repoDir, spawn, sourceRoot);
+    } finally {
+      console.log = originalLog;
+    }
+
+    assert.ok(logs.some(line => line.includes('Step7 commit 成功但 push 失敗')));
+    assert.ok(logs.some(line => line.includes('pre-receive hook declined')));
+  });
 });
 
 describe('cloneRepo', () => {

app/main.js

@@ -4,7 +4,7 @@ import { loadRoles, getRoleIntro } from './roles.js';
 import { getPRDiff, postComment } from './gitea.js';
 import { analyzeWithRole, loadOldFindings, mergeFindings, sortByLevel, deduplicateWithAI, loadExclusions, applyExclusions, filterFalsePositivesWithAI } from './findings.js';
 import { saveFindings, postOldFindingsComment, postNewNonCriticalComment, postNewCriticalComments } from './comments.js';
-import { cloneRepo, commitAndPush } from './git.js';
+import { cloneRepo, commitAndPush, getRepoState } from './git.js';
 import { validateJSONArrayFile, ensureJSONArrayFileExists } from './json.js';
 
 const WORKSPACE = process.env.GITHUB_WORKSPACE || '/workspace';
@@ -69,6 +69,10 @@ async function main() {
   } catch (e) {
     console.log(`  ⚠️  clone repo 失敗（繼續執行）: ${e.message}`);
   }
+  const repoState = repoDir ? getRepoState(repoDir) : null;
+  if (repoState) {
+    console.log(`  repo 狀態: branch=${repoState.branch || 'detached'} commit=${repoState.shortSha || 'unknown'} commit_time=${repoState.commitTime || 'unknown'} path=${repoState.repoDir}`);
+  }
   const oldFindings = loadOldFindings(repoDir || WORKSPACE);
   const mergedFindings = mergeFindings(oldFindings, newFindings);
   console.log(`  Step3 merged findings total=${mergedFindings.length}`);
@@ -81,14 +85,15 @@ async function main() {
   // Step5: 讀取排除問題檔案，過濾 PR 問題表格，並請 AI 判斷誤報
   console.log('\n🚫 Step4: AI 排除問題過濾');
   // 輸入至 findings 用於 AI 誤報過濾，exclusions 同時作為已知誤報參考
-  const exclusions = loadExclusions(repoDir || WORKSPACE);
+  const exclusions = loadExclusions(repoDir || WORKSPACE, repoState);
   const ruleFiltered = applyExclusions(sorted, exclusions);
   const filtered = await filterFalsePositivesWithAI(ruleFiltered, exclusions);
   console.log(`  Step4 完成: findings total=${filtered.length}`);
 
   // Step6: 寫入 findings.json，依序發布 comment
   console.log('\n📝 Step5: Findings 寫入與 Comment 發布');
-  saveFindings(WORKSPACE, filtered);
+  const reviewDir = repoDir || WORKSPACE;
+  saveFindings(WORKSPACE, filtered, reviewDir);
   try {
     await postOldFindingsComment(filtered);
     await postNewNonCriticalComment(filtered);
@@ -102,7 +107,7 @@ async function main() {
   console.log('\n🔎 Step6: JSON 格式驗證');
   const missingPaths = [];
   for (const relPath of [FINDINGS_PATH, EXCLUSIONS_PATH]) {
-    const fullPath = path.join(repoDir || WORKSPACE, relPath);
+    const fullPath = path.join(reviewDir, relPath);
     try {
       const result = await validateJSONArrayFile(fullPath, relPath);
       if (!result.exists) missingPaths.push({ fullPath, relPath });
@@ -117,7 +122,7 @@ async function main() {
 
   // Step7: commit/push findings.json 到來源分支
   console.log('\n💾 Step7: 記憶區 Commit/Push');
-  await commitAndPush(WORKSPACE, repoDir);
+  await commitAndPush(WORKSPACE, repoDir || WORKSPACE);
 
   // Step9: 有 critical 問題則 exit 1
   console.log('\n🚦 Step8: 嚴重問題檢查');