diff --git a/.github/aicodingflow-tests/test_product_docs_sync.py b/.github/aicodingflow-tests/test_product_docs_sync.py
index 348aa1bd..69ece102 100644
--- a/.github/aicodingflow-tests/test_product_docs_sync.py
+++ b/.github/aicodingflow-tests/test_product_docs_sync.py
@@ -342,6 +342,57 @@ def test_pr_body_includes_decision_source_pr_and_processed_entries(self) -> None
         self.assertIn("PR #86: Add first flow", body)
         self.assertIn("Document first flow.", body)
 
+    def test_pr_body_truncates_large_processed_decision_history(self) -> None:
+        entries = []
+        for number in range(1, 35):
+            entries.append(
+                {
+                    "pr": number,
+                    "title": f"PR {number}",
+                    "url": f"https://example.test/pull/{number}",
+                    "merged_at": f"2026-05-25T01:{number:02d}:00Z",
+                    "docs_update": "required",
+                    "reason": "Changed product behavior. " * 400,
+                    "affected_docs": [f"docs/product/raw/{number}.md"],
+                    "proposed_patch": "Document the product behavior in detail. " * 400,
+                }
+            )
+
+        body = body_writer.build_body(
+            pr_number="99",
+            pr_url="https://example.test/pull/99",
+            result={
+                "docs_update": "required",
+                "reason": "Product behavior changed.",
+                "affected_docs": ["docs/product/raw/flow.md"],
+                "source_context": ["PR #99"],
+                "proposed_patch": "Document the flow.",
+            },
+            ledger={"version": 1, "entries": entries},
+        )
+
+        self.assertLessEqual(len(body), body_writer.MAX_GITHUB_MARKDOWN_CHARS)
+        self.assertIn("Omitted 14 older processed decisions", body)
+        self.assertNotIn("PR #1: PR 1", body)
+        self.assertIn("PR #34: PR 34", body)
+        self.assertIn("Truncated change summary", body)
+
+    def test_pr_comment_truncates_large_patch_summary(self) -> None:
+        comment = body_writer.build_comment(
+            pr_number="87",
+            pr_url="https://example.test/pull/87",
+            result={
+                "docs_update": "required",
+                "reason": "Docs changed.",
+                "affected_docs": ["docs/product/raw/flow.md"],
+                "source_context": ["PR #87"],
+                "proposed_patch": "Long patch summary. " * 1000,
+            },
+        )
+
+        self.assertLessEqual(len(comment), body_writer.MAX_GITHUB_MARKDOWN_CHARS)
+        self.assertIn("Truncated patch summary", comment)
+
     def test_pr_comment_includes_current_run_decision(self) -> None:
         comment = body_writer.build_comment(
             pr_number="87",
diff --git a/.github/scripts/write_product_docs_sync_pr_body.py b/.github/scripts/write_product_docs_sync_pr_body.py
index 9f3132e6..545b6790 100644
--- a/.github/scripts/write_product_docs_sync_pr_body.py
+++ b/.github/scripts/write_product_docs_sync_pr_body.py
@@ -11,6 +11,25 @@
 
 
 DEFAULT_LEDGER_PATH = "docs/product/.product-docs-sync-ledger.json"
+MAX_GITHUB_MARKDOWN_CHARS = 60_000
+MAX_FIELD_CHARS = 4_000
+MAX_LEDGER_FIELD_CHARS = 800
+MAX_LEDGER_ENTRIES_IN_BODY = 20
+
+
+def truncate_text(value: Any, max_chars: int, *, label: str = "content") -> str:
+    text = str(value or "").strip()
+    if len(text) <= max_chars:
+        return text
+    suffix = f"\n\n[Truncated {label}; full content is available in workflow artifacts.]"
+    return text[: max(0, max_chars - len(suffix))].rstrip() + suffix
+
+
+def enforce_markdown_limit(markdown: str, max_chars: int = MAX_GITHUB_MARKDOWN_CHARS) -> str:
+    if len(markdown) <= max_chars:
+        return markdown
+    suffix = "\n\n[Truncated to stay within GitHub body length limits. See workflow artifacts for full details.]\n"
+    return markdown[: max(0, max_chars - len(suffix))].rstrip() + suffix
 
 
 def load_ledger(path: Path) -> dict[str, Any]:
@@ -41,12 +60,16 @@ def format_decision(entry: dict[str, Any]) -> list[str]:
     lines = [
         heading,
         f"  - docs update: `{entry.get('docs_update')}`",
-        f"  - reason: {entry.get('reason') or ''}",
+        f"  - reason: {truncate_text(entry.get('reason') or '', MAX_LEDGER_FIELD_CHARS, label='reason')}",
         f"  - url: {entry.get('url') or ''}",
         "  - affected docs:",
         *format_list(entry.get("affected_docs") or []),
     ]
-    proposed_patch = str(entry.get("proposed_patch") or "").strip()
+    proposed_patch = truncate_text(
+        entry.get("proposed_patch") or "",
+        MAX_LEDGER_FIELD_CHARS,
+        label="change summary",
+    )
     if proposed_patch:
         lines.extend(["  - change summary:", *[f"    {line}" for line in proposed_patch.splitlines()]])
     return lines
@@ -61,14 +84,21 @@ def build_body(pr_number: str, pr_url: str, result: dict[str, Any], ledger: dict
     affected_docs = result.get("affected_docs") or []
     source_context = result.get("source_context") or []
     processed_entries = ledger_entries(ledger or {})
-    return "\n".join(
+    shown_entries = processed_entries[-MAX_LEDGER_ENTRIES_IN_BODY:]
+    omitted_count = max(0, len(processed_entries) - len(shown_entries))
+    omitted_lines = (
+        [f"- Omitted {omitted_count} older processed decisions to keep the PR body within GitHub limits.", ""]
+        if omitted_count
+        else []
+    )
+    body = "\n".join(
         [
             "Synchronizes long-term product docs from merged implementation pull requests.",
             "",
             "Latest decision:",
             f"- source PR: #{pr_number}",
             f"- docs update: `{result.get('docs_update')}`",
-            f"- reason: {result.get('reason')}",
+            f"- reason: {truncate_text(result.get('reason'), MAX_FIELD_CHARS, label='reason')}",
             f"- source URL: {pr_url}",
             "",
             "Affected docs:",
@@ -78,18 +108,20 @@ def build_body(pr_number: str, pr_url: str, result: dict[str, Any], ledger: dict
             *(f"- {item}" for item in source_context),
             "",
             "Patch summary:",
-            str(result.get("proposed_patch") or ""),
+            truncate_text(result.get("proposed_patch"), MAX_FIELD_CHARS, label="patch summary"),
             "",
             "Processed decisions in this PR:",
+            *omitted_lines,
             *(
                 line
-                for entry in processed_entries
+                for entry in shown_entries
                 for line in [*format_decision(entry), ""]
             ),
             "This PR may accumulate multiple product docs sync decisions until it is reviewed and merged.",
             "",
         ]
     )
+    return enforce_markdown_limit(body)
 
 
 def build_comment(pr_number: str, pr_url: str, result: dict[str, Any]) -> str:
@@ -99,7 +131,7 @@ def build_comment(pr_number: str, pr_url: str, result: dict[str, Any]) -> str:
         "",
         f"- source PR: #{pr_number}",
         f"- docs update: `{result.get('docs_update')}`",
-        f"- reason: {result.get('reason') or ''}",
+        f"- reason: {truncate_text(result.get('reason'), MAX_FIELD_CHARS, label='reason')}",
         f"- source URL: {pr_url}",
         "",
         "Affected docs:",
@@ -108,12 +140,12 @@ def build_comment(pr_number: str, pr_url: str, result: dict[str, Any]) -> str:
     if not affected_docs:
         lines.append("- none")
 
-    proposed_patch = str(result.get("proposed_patch") or "").strip()
+    proposed_patch = truncate_text(result.get("proposed_patch") or "", MAX_FIELD_CHARS, label="patch summary")
     lines.extend(["", "Patch summary:", proposed_patch or "None."])
     if result.get("docs_update") == "uncertain":
         lines.extend(["", "This docs update is uncertain and needs maintainer confirmation."])
     lines.append("")
-    return "\n".join(lines)
+    return enforce_markdown_limit("\n".join(lines))
 
 
 def main() -> int:
diff --git a/docs/product/raw/product-docs-sync-workflow.md b/docs/product/raw/product-docs-sync-workflow.md
index 2a8e719d..2ba8d58c 100644
--- a/docs/product/raw/product-docs-sync-workflow.md
+++ b/docs/product/raw/product-docs-sync-workflow.md
@@ -91,6 +91,10 @@ PR comment，记录本次 run 的 source PR、`docs_update` 决策、原因、
 如果最新决策是 `uncertain`，追加 comment 也应明确提示需要维护者确认。workflow 不会编辑旧的
 bot comment 来替代追加记录。
 
+生成 PR body 和 comment 时，workflow 会保守控制输出长度，避免超过 GitHub body/comment
+长度限制。最新同步决策会保留较完整摘要；历史 ledger 决策只展示最近一批，并对过长的 reason
+或 patch summary 做截断，完整上下文仍可从 workflow artifacts 和 ledger 文件中追溯。
+
 长期产品文档只有在同步 PR 经过 review 并合并后才成为权威产品知识。
 
-来源：PR #179，https://github.com/Terry-Mao/AICodingFlow/pull/179；PR #184，https://github.com/Terry-Mao/AICodingFlow/pull/184；PR #187，https://github.com/Terry-Mao/AICodingFlow/pull/187；PR #188，https://github.com/Terry-Mao/AICodingFlow/pull/188；PR #194，https://github.com/Terry-Mao/AICodingFlow/pull/194；PR #215，Issue #214；PR #217，Issue #216。
+来源：PR #179，https://github.com/Terry-Mao/AICodingFlow/pull/179；PR #184，https://github.com/Terry-Mao/AICodingFlow/pull/184；PR #187，https://github.com/Terry-Mao/AICodingFlow/pull/187；PR #188，https://github.com/Terry-Mao/AICodingFlow/pull/188；PR #194，https://github.com/Terry-Mao/AICodingFlow/pull/194；PR #215，Issue #214；PR #217，Issue #216；Issue #257。
diff --git a/docs/product/wiki/concepts/product-docs-sync-workflow.md b/docs/product/wiki/concepts/product-docs-sync-workflow.md
index 9042344e..ee4f7f7c 100644
--- a/docs/product/wiki/concepts/product-docs-sync-workflow.md
+++ b/docs/product/wiki/concepts/product-docs-sync-workflow.md
@@ -57,6 +57,7 @@ Agent 写入 `product-docs-sync-result.json`，其中 `docs_update` 只能是：
 - 普通同步 PR title 是 `Update product docs`；最新决策为 `uncertain` 时使用 draft PR 和 `Draft: Update product docs` title。
 - `not-needed` 不修改权威 markdown 文档，但 ledger 更新仍会创建或更新只记录同步决策的 PR。
 - 每次创建或更新同步 PR 后，workflow 都会追加一条 conversation comment 记录本次 run；旧 bot comment 不会被编辑替代。
+- PR body 和 comment 会保守控制长度；历史 ledger 决策只展示最近一批，过长字段会被截断，完整上下文保留在 workflow artifacts 和 ledger 中。
 - 长期产品文档只有同步 PR 经 review 并合并后才成为权威知识。
 
 ## Related Concepts
diff --git a/docs/product/wiki/log.md b/docs/product/wiki/log.md
index abc24935..d3deacf5 100644
--- a/docs/product/wiki/log.md
+++ b/docs/product/wiki/log.md
@@ -1,5 +1,19 @@
 # Compile Log
 
+## 2026-06-10
+
+按 `docs/product/raw/` 作为权威来源同步 Product LLM Wiki 的产品文档同步行为：
+
+- 更新 [产品文档同步 workflow 摘要](summaries/product-docs-sync-workflow.md) 与 [产品文档同步 workflow](concepts/product-docs-sync-workflow.md)：记录产品文档同步 PR body/comment 的长度保护、历史 ledger 摘要和超长字段截断规则。
+
+## 待确认 / 开放问题
+
+- 待确认：raw source 未详细描述 spec agent 的全部输出文件契约，只说明外层 workflow 创建或更新 spec PR。
+
+## 冲突
+
+- 未发现 raw source 之间的明确冲突。
+
 ## 2026-06-09
 
 按 `docs/product/raw/` 作为权威来源复核 Product LLM Wiki，并校准近期 raw source 中已经确认的新规则：
diff --git a/docs/product/wiki/summaries/product-docs-sync-workflow.md b/docs/product/wiki/summaries/product-docs-sync-workflow.md
index 65bb14be..983eaf5d 100644
--- a/docs/product/wiki/summaries/product-docs-sync-workflow.md
+++ b/docs/product/wiki/summaries/product-docs-sync-workflow.md
@@ -45,6 +45,7 @@ Source: [docs/product/raw/product-docs-sync-workflow.md](../../raw/product-docs-
 - 普通同步 PR title 是 `Update product docs`；最新决策为 `uncertain` 时使用 draft PR 和 `Draft: Update product docs` title。
 - `not-needed` 不修改权威 markdown 文档，但 ledger 更新仍会创建或更新只记录同步决策的 PR。
 - 每次创建或更新同步 PR 后，workflow 都会追加一条 conversation comment，记录本次 run 的 source PR、决策、原因、受影响文档和 patch summary；旧 bot comment 不会被编辑替代。
+- PR body 和 comment 会保守控制长度；历史 ledger 决策只展示最近一批，过长 reason 或 patch summary 会被截断，完整上下文保留在 workflow artifacts 和 ledger 中。
 - 长期产品文档只有在同步 PR review 并合并后才成为权威产品知识。
 
 ## 支持的概念