diff --git a/models/claude-fable-5-cc-high.json b/models/claude-fable-5-cc-high.json
index d9c0fc4..9c6de1e 100644
--- a/models/claude-fable-5-cc-high.json
+++ b/models/claude-fable-5-cc-high.json
@@ -15,8 +15,8 @@
     },
     "turf-war": {
       "note": {
-        "zh": "Claude Fable 5(High 思考档位)在 Claude Code 中用仓库标准中文提示词(仅按案例规则填写文末【本次运行】占位符)一次生成的圈地智能体,未经人工修改。按案例规则,策略文件不公开、仅注册运行,由维护者亲自执行。自命名「疾风圈地」。",
-        "en": "One-shot Turf War agent by Claude Fable 5 (High effort) in Claude Code from the repository's canonical Chinese prompt (only the [This run] placeholders filled per case rules), with no human edits. Per case rules the policy file is unpublished; this entry registers the maintainer-executed run. Self-named \"疾风圈地\"."
+        "zh": "Claude Fable 5(High 思考档位)在 Claude Code 中用仓库标准中文提示词(仅按案例规则填写文末【本次运行】占位符)一次生成的圈地智能体,未经人工修改。按案例规则策略文件不公开、仅注册运行,由维护者亲自执行,任何 bug 均不修复。自命名「疾风圈地」。诊断:策略翻车,龟缩到极致:圈了 164 次却每次平均仅 3.9 格,勤劳画小圈、画一个被对手大圈盖一个,0 死 0 杀,终局仅 165 分。与 Opus xhigh 同病更甚——名叫疾风,啃得太碎。",
+        "en": "One-shot Turf War agent by Claude Fable 5 (High effort) in Claude Code from the repository's canonical Chinese prompt (only the [This run] placeholders filled per case rules), with no human edits. Per case rules the policy file is unpublished and no bug is ever fixed; this registers the maintainer-executed run. Self-named \"疾风圈地\". Diagnosis: Strategic failure, turtling to the extreme: 164 captures averaging just 3.9 cells each — swift tiny loops instantly overwritten by the opponent's bigger ones, 0 deaths and 0 kills, ending at only 165. The same trap as Opus xhigh but worse — swift name, bites too small."
       },
       "contributor": "Nagi-ovo",
       "agentName": "疾风圈地"
diff --git a/models/claude-opus-4-8-cc-xhigh.json b/models/claude-opus-4-8-cc-xhigh.json
index 83d844c..f673e9a 100644
--- a/models/claude-opus-4-8-cc-xhigh.json
+++ b/models/claude-opus-4-8-cc-xhigh.json
@@ -7,8 +7,8 @@
   "runs": {
     "turf-war": {
       "note": {
-        "zh": "Claude Opus 4.8(xhigh 思考档位)在 Claude Code 中用仓库标准中文提示词(仅按案例规则填写文末【本次运行】占位符)一次生成的圈地智能体,未经人工修改。按案例规则,策略文件不公开、仅注册运行,由维护者亲自执行。自命名「蚕食」。",
-        "en": "One-shot Turf War agent by Claude Opus 4.8 (xhigh effort) in Claude Code from the repository's canonical Chinese prompt (only the [This run] placeholders filled per case rules), with no human edits. Per case rules the policy file is unpublished; this entry registers the maintainer-executed run. Self-named \"蚕食\"."
+        "zh": "Claude Opus 4.8(xhigh 思考档位)在 Claude Code 中用仓库标准中文提示词(仅按案例规则填写文末【本次运行】占位符)一次生成的圈地智能体,未经人工修改。按案例规则策略文件不公开、仅注册运行,由维护者亲自执行,任何 bug 均不修复。自命名「蚕食」。诊断:策略翻车,过度保守:全程 0 死亡,但每次只圈 7 格左右的小块,圈来的地被对手的大圈反复覆盖,累计吃 776 格终局只剩 277。撞上提示词警告的失败模式——只求不败拿不到高排名。",
+        "en": "One-shot Turf War agent by Claude Opus 4.8 (xhigh effort) in Claude Code from the repository's canonical Chinese prompt (only the [This run] placeholders filled per case rules), with no human edits. Per case rules the policy file is unpublished and no bug is ever fixed; this registers the maintainer-executed run. Self-named \"蚕食\". Diagnosis: Strategic failure, over-cautious: 0 deaths all match, but each capture is a tiny ~7-cell loop that the opponent's bigger loops keep overwriting — it gains 776 cells across the match yet ends with 277. Exactly the failure the prompt warns about: merely avoiding losing does not rank."
       },
       "contributor": "Nagi-ovo",
       "agentName": "蚕食"
diff --git a/models/claude-sonnet-5-cc-max.json b/models/claude-sonnet-5-cc-max.json
new file mode 100644
index 0000000..f892dcd
--- /dev/null
+++ b/models/claude-sonnet-5-cc-max.json
@@ -0,0 +1,17 @@
+{
+  "label": "Claude Sonnet 5",
+  "vendor": "Anthropic",
+  "harness": "Claude Code",
+  "effort": "Max",
+  "order": 12.5,
+  "runs": {
+    "turf-war": {
+      "note": {
+        "zh": "Claude Sonnet 5(Max 思考档位)在 Claude Code 中用仓库标准中文提示词(仅按案例规则填写文末【本次运行】占位符)一次生成的圈地智能体,未经人工修改。按案例规则策略文件不公开、仅注册运行,由维护者亲自执行,任何 bug 均不修复。自命名「卷地风」。诊断:全程 0 次成功圈地,反复送死 11 次——离开领地后撞上自己的尾巴/岩石/边界而亡,一个闭环都没画成。属于莽夫送死型:太急、路径规划没避开自身尾巴。",
+        "en": "One-shot Turf War agent by Claude Sonnet 5 (Max effort) in Claude Code from the repository's canonical Chinese prompt (only the [This run] placeholders filled per case rules), with no human edits. Per case rules the policy file is unpublished and no bug is ever fixed; this registers the maintainer-executed run. Self-named \"卷地风\". Diagnosis: Zero completed captures and 11 self-inflicted deaths — it leaves home and dies on its own trail / a rock / the border before ever closing a loop. A reckless death-loop: too eager, its pathing never avoids its own trail."
+      },
+      "contributor": "Nagi-ovo",
+      "agentName": "卷地风"
+    }
+  }
+}
\ No newline at end of file
diff --git a/models/gpt-5-6-terra-codex-cli-xhigh.json b/models/gpt-5-6-terra-codex-cli-xhigh.json
index ff7e86f..573877a 100644
--- a/models/gpt-5-6-terra-codex-cli-xhigh.json
+++ b/models/gpt-5-6-terra-codex-cli-xhigh.json
@@ -7,8 +7,8 @@
   "runs": {
     "turf-war": {
       "note": {
-        "zh": "GPT-5.6-Terra(xhigh 思考档位)在 Codex CLI 中用仓库标准中文提示词(仅按案例规则填写文末【本次运行】占位符)一次生成的圈地智能体,未经人工修改。按案例规则,策略文件不公开、仅注册运行,由维护者亲自执行。自命名「拓荒狐」。该产物的 act() 在部分局面抛异常,按规则退化为该 tick 直行,不做任何修复——这是其真实 one-shot 表现。",
-        "en": "One-shot Turf War agent by GPT-5.6-Terra (xhigh effort) in Codex CLI from the repository's canonical Chinese prompt (only the [This run] placeholders filled per case rules), with no human edits. Per case rules the policy file is unpublished; this entry registers the maintainer-executed run. Self-named \"拓荒狐\" (Pioneer Fox). Its act() throws in some states and degrades to straight moves for those ticks per the rules; no fix applied — that is its true one-shot behavior."
+        "zh": "GPT-5.6-Terra(xhigh 思考档位)在 Codex CLI 中用仓库标准中文提示词(仅按案例规则填写文末【本次运行】占位符)一次生成的圈地智能体,未经人工修改。按案例规则策略文件不公开、仅注册运行,由维护者亲自执行,任何 bug 均不修复。自命名「拓荒狐」。诊断:工程翻车:自写的 addTop 插入排序有 off-by-one,数组留下 undefined 空洞,act() 几乎每个 tick 都抛 TypeError(top[at-1].raw)。按规则抛异常即该 tick 直行,于是它全程走直线、策略一步未跑成。这是它真实的一次成型水平。",
+        "en": "One-shot Turf War agent by GPT-5.6-Terra (xhigh effort) in Codex CLI from the repository's canonical Chinese prompt (only the [This run] placeholders filled per case rules), with no human edits. Per case rules the policy file is unpublished and no bug is ever fixed; this registers the maintainer-executed run. Self-named \"拓荒狐\". Diagnosis: Engineering failure: its own addTop insertion-sort has an off-by-one that leaves an undefined hole, so act() throws a TypeError (top[at-1].raw) almost every tick. Per the rules a thrown act() means a straight move, so it walks in a straight line all match — its real strategy never runs once."
       },
       "contributor": "Nagi-ovo",
       "agentName": "拓荒狐"