diff --git a/assets/expert-panel/icon.png b/assets/expert-panel/icon.png
new file mode 100644
index 0000000..56f3d89
Binary files /dev/null and b/assets/expert-panel/icon.png differ
diff --git a/external_plugins/expert-panel/.zcode-plugin/plugin.json b/external_plugins/expert-panel/.zcode-plugin/plugin.json
new file mode 100644
index 0000000..0f44ceb
--- /dev/null
+++ b/external_plugins/expert-panel/.zcode-plugin/plugin.json
@@ -0,0 +1,13 @@
+{
+  "name": "expert-panel",
+  "version": "0.1.2",
+  "description": "专家团共识评审：多位不同视角专家独立作答→真匿名互评→仲裁归纳（共识率脚本计算、不足自动加辩）→主席综合→事实复核；quick/standard/deep 三档，零配置，结论按一致/分歧分离绝不硬合成。",
+  "description_i18n": {
+    "en": "Expert-panel consensus review: independent answers from multiple expert lenses → truly anonymous cross-critique → arbitration (consensus rate computed by script, extra debate round below threshold) → chair synthesis → fact verification; quick/standard/deep profiles, zero config, conclusions separate agreed items from explicit disagreements.",
+    "zh-CN": "专家团共识评审：多位不同视角专家独立作答→真匿名互评→仲裁归纳（共识率脚本计算、不足自动加辩）→主席综合→事实复核；quick/standard/deep 三档，零配置，结论按一致/分歧分离绝不硬合成。"
+  },
+  "author": { "name": "Zhang24y" },
+  "license": "MIT",
+  "skills": "./skills",
+  "commands": "./commands"
+}
diff --git a/external_plugins/expert-panel/LICENSE b/external_plugins/expert-panel/LICENSE
new file mode 100644
index 0000000..e883e6f
--- /dev/null
+++ b/external_plugins/expert-panel/LICENSE
@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 Zhang24y
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.
diff --git a/external_plugins/expert-panel/README.md b/external_plugins/expert-panel/README.md
new file mode 100644
index 0000000..da2576c
--- /dev/null
+++ b/external_plugins/expert-panel/README.md
@@ -0,0 +1,101 @@
+# expert-panel — Expert Panel Consensus Review (ZCode Plugin)
+
+Put an important decision in front of a panel of experts with distinct lenses: **independent answering → truly anonymous cross-critique → arbitration (consensus rate computed deterministically by script, extra debate round when below threshold) → chair synthesis → fact verification**. The output separates **stable agreed conclusions from explicitly stated disagreements**. Zero configuration — every role runs on your current session model.
+
+## Why
+
+The failure modes of a single model answer: conformity (anchoring on the first framing), authority bias (deferring to whoever sounds senior), and forced synthesis (mushing disagreement into "generally feasible"). This workflow counters each one:
+
+| Problem | Countermeasure |
+|---|---|
+| Conformity | Every expert **answers independently first**, before seeing anyone else's position |
+| Authority bias | Critique inputs contain only anonymous labels (Expert A/B/C) + position + key points — **no lens name, no model** |
+| False consensus | Consensus rate = agreed/(agreed+disputed), **computed deterministically by script — never self-reported by a model** |
+| Splitting the difference | Below threshold → one extra debate round; still below → **disagreements are stated explicitly for human adjudication, never force-merged** |
+| Fabricated facts | Objectively verifiable agreed items are checked by a fact verifier (it reads files / runs commands when it can); failures are dropped |
+
+A single expert failing is tolerated; the conclusion is annotated with sample completeness (absent lenses / valid expert count).
+
+## Installation
+
+**Local directory marketplace** (put this repo anywhere on your machine):
+
+1. ZCode → Plugin Marketplace → Add → Add Plugin Marketplace, paste the repo **directory** path
+2. Personal → the marketplace → expert-panel → Install
+3. Takes effect in a new session
+
+**CLI** (after the marketplace is added):
+
+```bash
+zcode plugins install expert-panel@<marketplace-name>
+```
+
+Requirements: ZCode desktop (the dynamic-workflows runtime is bundled). The run ledger uses `python3` if present (missing python3 does not block conclusions; the run is just not logged).
+
+## Usage
+
+Three ways to trigger it once installed:
+
+- Natural language: "use the expert panel to evaluate X", "have the expert panel settle this", "give this decision a multi-perspective check"
+- Slash command: `/expert-panel <your question>`
+- Direct call: `CreateWorkflow(path=<the plugin's workflows/expert-panel.dwf.ts>, args={...})`
+
+### Parameters
+
+| Parameter | Type | Default | Description |
+|---|---|---|---|
+| question | string | required | The question/decision the panel should settle |
+| profile | string | standard | quick (3 experts/1 round, minutes) \| standard (2 rounds) \| deep (auto-adds contrarian + synthesizer lenses, 2 rounds) |
+| experts | json | [务实工程师, 风险与边界审查, 产品价值] | Expert lens names (no duplicates) |
+| threshold | number | 0.6 | Consensus-rate threshold; below it one extra debate round runs |
+| max_rounds | number | quick=1, else 2 | Max rounds including the first |
+| evidence | json | empty | Evidence pack: texts or file paths (experts/verifier read files with their own tools) |
+| basis | json | empty | Review rubric: file paths or entries; when given, verdicts must cite the rubric item by item |
+| ledger_path | string | ~/.expert-panel/ledger.jsonl | Run ledger JSONL path |
+| simulate_failures | json | empty | Test-only: indexes of experts to simulate as failed |
+
+### Model selection
+
+By default every role uses the session default model — this is deliberate: **zero config, runs on any ZCode environment**. To switch models uniformly, pass `subagent_model` on the `CreateWorkflow` call (look up ids with `ListModels` on the current provider). Per-role model routing is intentionally unsupported by the host: a subagent's model must be known before the run starts.
+
+### Interpreting results
+
+- `agreementRate ≥ threshold`: consensus reached. `agreed` holds verified stable conclusions; `unverifiable` holds value judgements (cannot be objectively verified — they are the panel's opinion only).
+- `< threshold`: `disputed` lists each side's arguments for human adjudication — **this is a feature, not a failure**.
+- `absent` non-empty or fewer than 2 valid experts: the conclusion is annotated with sample completeness; mind the caveat when quoting.
+
+Each run appends one ledger line (`ledger_path`): question digest / profile / expert counts / rounds / consensus rate / agreed·disputed·rejected counts / stable flag.
+
+### 30-second verification
+
+In a new session, send:
+
+```
+Use the expert panel, quick profile, to evaluate: Markdown files vs a notes app for a personal knowledge base?
+```
+
+Expected (within minutes): an "Expert Panel Conclusion" card with a consensus rate (e.g. 74%), agreed items (with verification marks), disputed items (if any), an executive summary and a final recommendation; while running, the expert board advances from "independent answering" to "converged". A rate below 60% triggers an extra debate round and explicit disagreements — that is also normal behavior.
+
+## Data behavior & uninstall
+
+- Each run appends one JSONL line to `~/.expert-panel/ledger.jsonl` (relocate via `ledger_path`); beyond that it **writes nothing** — no hooks, no MCP server, no outbound network requests (review subagents only use your current session's own model channel).
+- Uninstall: Settings → Plugins → expert-panel → Uninstall. The only residue is `~/.expert-panel/` (the ledger dir); delete it to remove everything.
+
+## Design notes
+
+- **Anonymity protocol**: critique inputs carry only anonymous labels/positions/key points; lens names exist only inside each expert's own system prompt during independent answering.
+- **Deterministic consensus rate**: the script computes `agreed/(agreed+disputed)`; round gating keys off that value alone — a model cannot shade the arithmetic.
+- **Deep-profile guardrail**: a contrarian lens is mandatory (to counter "debate amplifying biased consensus"); a synthesizer is added when headcount is short.
+- **Origins**: the mechanism draws on the anonymous cross-critique and chair ideas of karpathy/llm-council, plus the author's own expert-panel workflow iterated in production on ZCode (v2.1: rubric injection, single-expert fault tolerance, sample-completeness annotation). No code was copied from llm-council (which currently ships no license); this plugin is an independent implementation, MIT-licensed.
+
+## Changelog
+
+Version is whatever the marketplace shows (the source of truth is plugin.json).
+
+- 0.1.2: ledger schema v2 — dual consensus rates (rate_pre at arbitration time / rate_post recomputed after verification, fixing the old unrecomputable-rate bug), per-round convergence series, and threshold/question_id fields logged in full.
+- 0.1.1: fix ledger write for bare-filename custom paths; pre-release cleanup (neutral git identity).
+- 0.1.0: first releasable version, full six-stage mechanism.
+
+## License
+
+MIT — see [LICENSE](LICENSE)
diff --git a/external_plugins/expert-panel/README_CN.md b/external_plugins/expert-panel/README_CN.md
new file mode 100644
index 0000000..80035d8
--- /dev/null
+++ b/external_plugins/expert-panel/README_CN.md
@@ -0,0 +1,106 @@
+# expert-panel —— 专家团共识评审（ZCode 插件）
+
+把一个重要决策交给多位不同视角的专家：**独立作答 → 真匿名互评 → 仲裁归纳 → 共识率门控加辩 → 主席综合 → 事实复核**，产出「稳定一致项 + 明示分歧」的结构化结论。零配置，装完即用，所有角色跑在你当前的会话模型上。
+
+## 为什么需要它
+
+单个模型一次回答的问题：从众（先入为主）、权威偏置（谁名气大听谁的）、硬合成（把分歧和稀泥成"总体可行"）。本工作流逐条对治：
+
+| 问题 | 对策 |
+|---|---|
+| 从众 | 每位专家**先独立作答**，之后才看到别人的立场 |
+| 权威偏置 | 互评输入只有「专家A/B/C」匿名标签 + 立场 + 要点，**不含视角名与模型** |
+| 虚假共识 | 共识率 = 一致项/(一致项+分歧项)，**由脚本确定性计算，不信模型自报** |
+| 和稀泥 | 共识率低于阈值自动加辩一轮；仍不足则**明示分歧交人工裁决，绝不硬合成** |
+| 编造事实 | 可客观验证的一致项由事实复核员核验（能读文件/跑命令就去验），不实剔除 |
+
+单专家失败自动容错，结论标注样本完整度（缺席视角/有效专家数）。
+
+## 安装
+
+**本地目录市场**（先把本仓库放到本机任意目录）：
+
+1. ZCode → 插件市场 → 添加 → 添加插件市场，粘贴本仓库所在**目录**路径
+2. 个人 → 该市场 → expert-panel → 安装
+3. 新会话生效
+
+**CLI**（已添加市场后）：
+
+```bash
+zcode plugins install expert-panel@<市场名>
+```
+
+要求：ZCode 桌面版（自带 dynamic workflows 运行时）。台账落盘用到 `python3`（缺失不阻断结论，仅不入账）。
+
+## 使用
+
+装好后三种触发方式：
+
+- 自然语言：「用专家团评估一下 X」「让专家团给个结论」「这个问题帮我多视角把关」
+- 斜杠命令：`/expert-panel <你的问题>`
+- 直接调用：`CreateWorkflow(path=<插件内 workflows/expert-panel.dwf.ts>, args={...})`
+
+### 参数
+
+| 参数 | 类型 | 默认 | 说明 |
+|---|---|---|---|
+| question | string | 必填 | 要专家团给出稳定结论的问题/决策 |
+| profile | string | standard | quick（3 专家/1 轮，数分钟）\| standard（2 轮）\| deep（自动补反方质询+综合连接原型，2 轮） |
+| experts | json | [务实工程师, 风险与边界审查, 产品价值] | 专家视角名数组（勿重复） |
+| threshold | number | 0.6 | 共识率阈值，低于自动加辩一轮 |
+| max_rounds | number | quick=1，其余=2 | 最大轮数（含首轮） |
+| evidence | json | 空 | 证据包：文本或文件路径数组（专家/复核员用工具自行读取） |
+| basis | json | 空 | 评测依据：文件路径或条目数组，要求逐条对照判定并引用条目 |
+| ledger_path | string | ~/.expert-panel/ledger.jsonl | 运行台账 JSONL 落盘路径 |
+| simulate_failures | json | 空 | 测试用：模拟作答失败的专家下标 |
+
+### 模型选择
+
+默认所有角色用会话默认模型——这是刻意设计：**零配置、在任何 ZCode 环境都能跑**。要统一换模型，在 `CreateWorkflow` 调用上传 `subagent_model`（用 `ListModels` 查当前 provider 可用的 id）。按角色分模型（如互评用轻量模型）宿主刻意不支持：子代理模型必须在运行前已知。
+
+### 结果解读
+
+- `agreementRate ≥ threshold`：达成共识。`agreed` 为经复核的稳定结论；`unverifiable` 是价值判断类（无法客观复核，仅代表专家团意见）。
+- `< threshold`：`disputed` 已列明各方论据，交人工裁决——**这是特性不是失败**。
+- `absent` 非空或有效专家 < 2：结论已标注样本完整度，引用时注意口径。
+
+### 30 秒验证（装完即测）
+
+新会话里发一句：
+
+```
+用专家团 quick 档评估：个人知识库用 Markdown 文件还是笔记软件？
+```
+
+预期输出（数分钟内）：一张「专家团结论」卡片，含共识率百分比（如 74%）、一致项清单（带复核标记）、分歧项清单（若有）、执行摘要与最终建议；运行中可见专家状态看板从「独立作答」推进到「已收敛」。共识率低于 60% 时会自动加辩一轮并明示分歧——同样算工作正常。
+
+每次运行一行追加台账（`ledger_path`）：问题摘要 / 档位 / 专家数 / 有效数 / 轮数 / 共识率 / 一致·分歧·剔除计数 / 是否达成。
+
+## 输出形态
+
+运行中：专家团状态看板（独立作答→交叉互评→修订表态→已收敛）+ 共识率收敛曲线 + 指标条。
+结束后：主产物为一份结论卡片（一致项/分歧项/复核剔除/执行摘要/最终建议/仲裁附录），并返回结构化 JSON（供编排方程序化消费）。
+
+## 数据行为与卸载
+
+- 每次运行追加一行 JSONL 到 `~/.expert-panel/ledger.jsonl`（可用 `ledger_path` 改路径）；除此之外**不写任何文件**，无 hooks、无 MCP server、无出站网络请求（评审子代理只用你当前会话本身的模型通道）。
+- 卸载：Settings → Plugins → expert-panel → 卸载。残留物仅 `~/.expert-panel/`（台账目录），删除即彻底清除。
+
+## 设计说明
+
+- **匿名协议**：互评阶段输入仅含匿名标签/立场/要点；视角名只在独立作答阶段存在于各自的 system prompt 中。
+- **共识率确定性**：脚本按 `一致/(一致+分歧)` 计算，轮间比较与加辩门控全部基于该值，模型无法影响分母分子之外的东西。
+- **deep 档护栏**：强制包含反共识原型（Contrarian），对冲「辩论放大偏差共识」；人数不足再补综合连接（Synthesizer）。
+- **渊源**：机制蓝本来自 karpathy/llm-council 的匿名互评与主席制思想 + 本作者在 ZCode 上长期使用的 expert-panel 工作流实证迭代（v2.1：评测依据注入、单专家容错、样本完整度标注）。
+
+## Changelog
+
+版本以插件市场显示为准（事实源在 plugin.json）。
+
+- 0.1.2：台账 schema v2——双口径共识率（rate_pre 仲裁时点 / rate_post 复核后重算，修复旧版 rate 不可复算问题）、逐轮收敛序列、threshold/question_id 等字段全量入账。
+- 0.1.1：修复自定义 ledger_path 为裸文件名时无法落账的边界问题；开源前清理（git 历史中性身份）。
+- 0.1.0：首个可发布版本，六阶段机制全量移植。
+
+## License
+
+MIT（见 [LICENSE](LICENSE)）
diff --git a/external_plugins/expert-panel/commands/expert-panel.md b/external_plugins/expert-panel/commands/expert-panel.md
new file mode 100644
index 0000000..0d3a620
--- /dev/null
+++ b/external_plugins/expert-panel/commands/expert-panel.md
@@ -0,0 +1,8 @@
+---
+description: 把一个问题交给专家团共识评审（独立作答→匿名互评→仲裁→加辩→主席→复核）
+---
+
+按 expert-panel skill（本插件 `skills/expert-panel/SKILL.md`）的流程调用专家团工作流：
+
+- 用户消息中除指令外的内容就是要评审的问题（question）；用户点名了档位就用（quick/standard/deep），没点名按 skill 默认 standard，急活可用 quick。
+- 调用方式、完整参数表、结果解读一律以 skill 为准。
diff --git a/external_plugins/expert-panel/skills/expert-panel/SKILL.md b/external_plugins/expert-panel/skills/expert-panel/SKILL.md
new file mode 100644
index 0000000..99ce4f9
--- /dev/null
+++ b/external_plugins/expert-panel/skills/expert-panel/SKILL.md
@@ -0,0 +1,63 @@
+---
+name: expert-panel
+description: 专家团共识评审：把重要决策/方案交给多位不同视角专家——独立作答→真匿名互评→仲裁归纳（共识率脚本确定性计算、不足自动加辩）→主席综合→事实复核，产出「稳定一致项 + 明示分歧」的结构化结论。当用户要"更稳"的结论、说 用专家团评估X / 让专家团给结论 / 多视角评审 / 共识评审 / 重要决策把关 时使用；急活用 quick 档（数分钟）。
+---
+
+# expert-panel —— 专家团共识评审
+
+## 机制（为什么可信）
+
+1. **独立作答**：N 位不同视角专家先各自独立表态，互不知晓（无从众）。
+2. **真匿名互评**：互评输入只有「专家A/B/C」标签 + 立场 + 要点，不含视角名（防权威偏置）；各自提出反对/补充并按说服力排序。
+3. **仲裁归纳**：归纳为一致项/分歧项。**共识率 = 一致/(一致+分歧)，由脚本确定性计算，不信模型自报**。
+4. **共识率门控加辩**：低于阈值（默认 0.6）自动加辩一轮，仍不足则**明示分歧交人工裁决，绝不硬合成**。
+5. **主席综合**：执行摘要 + 最终建议。
+6. **事实复核**：可客观验证的一致项由复核员核验（能读文件/跑命令就去验），不实的剔除；价值判断类标 unverifiable 原样保留。
+
+单专家失败自动容错，结论标注样本完整度（缺席视角/有效专家数）。
+
+## 怎么调用
+
+工作流文件随本插件分发：**本 skill 目录的 `../../workflows/expert-panel.dwf.ts`**（相对本 SKILL.md）。
+
+用 `CreateWorkflow` 的 **path 方式**调用（不要复制进用户的 workflows 目录，升级由插件统一管）：
+
+```
+CreateWorkflow(
+  path = <按上述相对位置解析出的绝对路径>,
+  name = "专家团",
+  args = { question, profile, ... }
+)
+```
+
+### 参数
+
+| 参数 | 类型 | 默认 | 说明 |
+|---|---|---|---|
+| question | string | 必填 | 要专家团给出稳定结论的问题/决策 |
+| profile | string | standard | quick（3 专家/1 轮，数分钟）\| standard（2 轮）\| deep（自动补反方质询+综合连接原型，2 轮） |
+| experts | json | [务实工程师, 风险与边界审查, 产品价值] | 专家视角名数组（勿重复） |
+| threshold | number | 0.6 | 共识率阈值，低于自动加辩一轮 |
+| max_rounds | number | quick=1，其余=2 | 最大轮数（含首轮） |
+| evidence | json | 空 | 证据包：文本或文件路径数组（专家/复核员用工具自行读取） |
+| basis | json | 空 | 评测依据：文件路径或条目数组，给了就要求逐条对照判定并引用条目 |
+| ledger_path | string | ~/.expert-panel/ledger.jsonl | 运行台账 JSONL 落盘路径 |
+| simulate_failures | json | 空 | 测试用：要模拟作答失败的专家下标数组 |
+
+### 模型选择（可选）
+
+所有角色默认用**会话默认模型，零配置可跑**。要统一换模型：在 `CreateWorkflow` 调用上传 `subagent_model`（id 用 `ListModels` 查当前 provider 的行）。
+注意：按角色分模型（如互评用轻量模型）是宿主刻意不支持的——子代理模型必须在运行前已知；请勿尝试在脚本里做运行时路由。
+
+### 结果解读
+
+- `agreementRate ≥ threshold`：达成共识，`agreed` 为经复核的稳定结论；`unverifiable` 是价值判断类。
+- `< threshold`：`disputed` 已列明各方论据，交人工裁决——这是特性不是失败。
+- `absent` 非空或有效专家 < 2：结论已标注样本完整度，引用时注意口径。
+- 每轮运行一行追加台账（`ledger_path`）：问题摘要/档位/专家数/轮数/共识率/一致分歧计数。
+
+## 边界
+
+- 价值判断类结论无法客观复核（标 unverifiable），仅代表专家团意见。
+- 台账写入失败不阻断结论（结果 `verified` 里会注明未入账）。
+- 深档会显著增加 token 消耗，重要决策再用。
diff --git a/external_plugins/expert-panel/workflows/expert-panel.dwf.ts b/external_plugins/expert-panel/workflows/expert-panel.dwf.ts
new file mode 100644
index 0000000..f856ec9
--- /dev/null
+++ b/external_plugins/expert-panel/workflows/expert-panel.dwf.ts
@@ -0,0 +1,427 @@
+/* zcode-workflow
+description: 专家团共识工作流：N 位不同视角的专家独立作答 → 真匿名交叉互评（只见 专家A/B/C 标签，
+  不见视角名）→ 仲裁人归纳一致项/分歧项（共识率由脚本确定性计算，低于阈值自动加辩）→ 主席执行摘要 →
+  可验证事实复核。单专家失败自动容错并标注样本完整度；profile（quick/standard/deep）控制轮数与专家构成。
+  可移植版：所有角色用会话默认模型，零配置可跑；要统一换模型在调用层用 CreateWorkflow 的 subagent_model。
+  输出按「一致(稳定)/分歧(明示)」分离，防硬合成与偏见共识。
+whenToUse: 用户要"更稳"的结论/重要决策/方案评审/多视角评估时；说"用专家团评估 X / 让专家团给结论"。
+  急活用 profile=quick（数分钟），重要产出用 standard，最高把关用 deep。共识率不足自动加辩，
+  仍不足则明示分歧交人工裁决。
+args:
+  basis:
+    type: json
+    description: 评测依据，文件路径或条目数组；给了就要求专家与复核员逐条对照依据判定并引用所在条目（空数组=明确不用，默认不注入）
+  evidence:
+    type: json
+    description: 证据包，文本或文件路径数组（路径由专家/复核员用工具自行读取），默认空
+  experts:
+    type: json
+    description: 专家视角名数组（勿重复），默认 [务实工程师, 风险与边界审查, 产品价值]
+  ledger_path:
+    type: string
+    description: 运行台账 JSONL 落盘路径（每轮一行），默认 ~/.expert-panel/ledger.jsonl
+  max_rounds:
+    type: number
+    description: 最大轮数（含首轮），不填由 profile 决定（quick=1，其余=2）
+  profile:
+    type: string
+    description: quick（3专家/1轮）| standard（默认，2轮）| deep（自动补足反方质询与综合连接两类原型/2轮）
+    default: standard
+  question:
+    type: string
+    description: 要专家团给出稳定结论的问题/决策
+    required: true
+  question_id:
+    type: string
+    description: 题目标识（基准跑批/回溯归因用），默认空
+  simulate_failures:
+    type: json
+    description: 测试用，要模拟作答失败的专家下标数组，默认空
+  threshold:
+    type: number
+    description: 共识率阈值，低于则再辩一轮，仍低则明示分歧交人工裁决
+    default: 0.6
+*/
+interface ExpertAnswer {
+  /** 对问题的立场或方案（一句话）。 */
+  position: string;
+  /** 支撑要点，每条尽量独立可核。 */
+  keyPoints: string[];
+  /** 自评置信度 0-1。 */
+  confidence: number;
+}
+
+interface Objection {
+  /** 针对哪位匿名专家，填其标签（如 专家B）。 */
+  target: string;
+  /** 反对或补充意见。 */
+  comment: string;
+}
+
+interface Critique {
+  /** 对其他专家立场的反对与补充。 */
+  objections: Objection[];
+  /** 其他专家按说服力从高到低的标签排序（如 ["专家C","专家A"]）。 */
+  ranking: string[];
+}
+
+interface Revision {
+  /** 修订后的立场（可维持原判）。 */
+  position: string;
+  /** 修订后要点。 */
+  keyPoints: string[];
+  /** 是否改变了立场。 */
+  changed: boolean;
+}
+
+interface Dispute {
+  /** 分歧主题。 */
+  topic: string;
+  /** 各方论据摘要（用专家标签指代）。 */
+  sides: string;
+}
+
+interface Consensus {
+  /** 各专家均认可的结论——稳定输出。 */
+  agreed: string[];
+  /** 仍存分歧的结论。 */
+  disputed: Dispute[];
+  /** 仲裁人的最终建议。 */
+  recommendation: string;
+}
+
+interface Verification {
+  /** 复核后保留的一致项。 */
+  agreed: string[];
+  /** 复核未通过的项及原因。 */
+  rejected: { item: string; reason: string }[];
+  /** 价值判断类或无法客观复核的项，原样保留。 */
+  unverifiable: string[];
+}
+
+interface ChairSummary {
+  /** 3-5 句执行摘要：共识是什么、分歧在哪、怎么收敛的。 */
+  executiveSummary: string;
+  /** 面向用户的明确最终建议，含适用条件。 */
+  finalRecommendation: string;
+}
+
+/** 共识率由脚本确定性计算，不信仲裁人自报的数字。 */
+function computeRate(c: Consensus): number {
+  const a = c.agreed.length;
+  const d = c.disputed.length;
+  return a + d === 0 ? 0 : a / (a + d);
+}
+
+/** 专家下标 → 匿名标签（专家A/B/C…）。 */
+function labelOf(i: number): string {
+  return `专家${String.fromCharCode(65 + i)}`;
+}
+
+// ---------- 参数解析 ----------
+const question = String(args.question);
+const profileRaw = String(args.profile ?? "standard");
+const profile: "quick" | "standard" | "deep" = profileRaw === "quick" ? "quick" : profileRaw === "deep" ? "deep" : "standard";
+const defaultLenses = ["务实工程师", "风险与边界审查", "产品价值"];
+let lenses: string[] = Array.isArray(args.experts) && (args.experts as unknown[]).length > 0
+  ? (args.experts as unknown[]).map(String)
+  : defaultLenses;
+if (profile === "quick") lenses = lenses.slice(0, 3);
+if (profile === "deep") {
+  // 护栏：deep 档必须含反共识原型（对冲「辩论放大偏差共识」），人数不足再补综合型
+  if (!lenses.some((l) => /反方|反对|质询|质疑|Contrarian|魔鬼/i.test(l))) lenses = [...lenses, "反方质询（Contrarian）"];
+  if (lenses.length < 4) lenses = [...lenses, "综合连接（Synthesizer）"];
+}
+const threshold = typeof args.threshold === "number" && args.threshold > 0 && args.threshold <= 1 ? args.threshold : 0.6;
+const maxRounds = typeof args.max_rounds === "number" && args.max_rounds >= 1
+  ? Math.floor(args.max_rounds)
+  : profile === "quick" ? 1 : 2;
+const simulateFailures: number[] = Array.isArray(args.simulate_failures)
+  ? (args.simulate_failures as unknown[]).map(Number).filter((n) => Number.isInteger(n))
+  : [];
+const evidence: string[] = Array.isArray(args.evidence)
+  ? (args.evidence as unknown[]).map(String).filter((s) => s.trim().length > 0)
+  : [];
+const evidenceBlock = evidence.length > 0
+  ? `\n证据包（${evidence.length} 项；文件路径请用你的工具读取后再引用，结论须标注证据出处）：\n${evidence.map((e, i) => `[${i + 1}] ${e}`).join("\n")}`
+  : "";
+const basisGiven = Array.isArray(args.basis);
+const basis: string[] = basisGiven
+  ? (args.basis as unknown[]).map(String).filter((s) => s.trim().length > 0)
+  : [];
+const basisBlock = basis.length > 0
+  ? `\n评测依据（调用方指定；评审前用工具读取与问题相关的原文，结论须对照依据逐条判定并引用所在条目，不凭印象打分；与问题无关的条目可略过并说明）：\n${basis.map((b, i) => `[${i + 1}] ${b}`).join("\n")}`
+  : "";
+const ledgerPath = typeof args.ledger_path === "string" && args.ledger_path.trim().length > 0
+  ? args.ledger_path.trim()
+  : "~/.expert-panel/ledger.jsonl";
+
+// ---------- 仪表盘 ----------
+artifact.board("panel", {
+  title: "专家团状态",
+  key: "name",
+  status: "stage",
+  columns: ["独立作答", "交叉互评", "修订表态", "已收敛", "缺席"],
+  cardTitle: "name",
+  detail: [{ field: "position", label: "立场" }],
+});
+artifact.chart("consensus", {
+  title: "共识率收敛曲线",
+  x: { field: "round", label: "轮次" },
+  y: { field: "rate", label: "共识率" },
+  baseline: { field: "target", label: "阈值" },
+});
+artifact.metrics("stats", {
+  title: "专家团指标",
+  metrics: [
+    { field: "ratePct", label: "共识率" },
+    { field: "round", label: "轮次" },
+    { field: "alive", label: "有效专家" },
+    { field: "profileLabel", label: "档位" },
+  ],
+});
+
+const experts = lenses.map((lens) =>
+  agent(`专家·${lens}`, {
+    system:
+      `你是「${lens}」视角的资深专家。先独立给出立场、3-6 条支撑要点和置信度；` +
+      "只凭自己的专业判断，不要猜测别人会怎么说，也不要为了和谐改变有依据的立场；确实被说服才改。用中文回答。",
+  }),
+);
+
+phase("独立作答：每位专家独立给出立场与论据");
+// 仲裁需要全部有效专家的独立立场，此处必须等齐；单个专家失败不拖垮整体（allSettled）
+const settled = await Promise.allSettled(
+  lenses.map((lens, i) =>
+    simulateFailures.includes(i)
+      ? Promise.reject(new Error("模拟作答失败（测试注入）"))
+      : experts[i].ask<ExpertAnswer>(`问题：${question}\n视角：${lens}${evidenceBlock}${basisBlock}`),
+  ),
+);
+const active: number[] = [];
+const answers: ExpertAnswer[] = [];
+settled.forEach((s, i) => {
+  if (s.status === "fulfilled") {
+    active.push(i);
+    answers.push(s.value);
+    report({ name: lenses[i], stage: "独立作答", position: s.value.position }, "panel");
+  } else {
+    report({ name: lenses[i], stage: "缺席", position: `作答失败：${String(s.reason).slice(0, 80)}` }, "panel");
+  }
+});
+if (active.length === 0) throw new Error("所有专家均未完成作答，无法仲裁");
+const absent = lenses.filter((_, i) => !active.includes(i));
+log(`${active.length}/${lenses.length} 位专家完成独立作答${absent.length > 0 ? `（缺席：${absent.join("、")}）` : ""}`);
+
+/** 第 m 位有效专家能看到的他人立场：只有匿名标签、立场、要点——不含视角名与模型。 */
+function othersFor(m: number) {
+  return active
+    .map((idx, n) => ({ idx, n }))
+    .filter((o) => o.n !== m)
+    .map((o) => ({ 专家: labelOf(o.idx), position: answers[o.n].position, keyPoints: answers[o.n].keyPoints }));
+}
+
+phase("交叉互评：匿名互看立场，提出反对与补充");
+const critiqueSettled = await Promise.allSettled(
+  active.map((idx, m) =>
+    agent(`互评·${lenses[idx]}`, {
+      system:
+        `你是「${lenses[idx]}」视角的独立评审。只针对其他专家的匿名立场提出反对与补充，` +
+        "不重复自己的论点，就事论事。用中文回答。",
+    }).ask<Critique>(
+      `问题：${question}\n你此前的立场：${answers[m].position}\n` +
+        `其他专家的匿名立场：${JSON.stringify(othersFor(m))}\n` +
+        "请只针对他们的立场提出反对与补充（target 填对方标签），并给出 ranking（对方标签数组，说服力从高到低）。",
+    ),
+  ),
+);
+const critiques: Critique[] = critiqueSettled.map((s) => (s.status === "fulfilled" ? s.value : { objections: [], ranking: [] }));
+active.forEach((idx, m) => report({ name: lenses[idx], stage: "交叉互评", position: answers[m].position }, "panel"));
+
+phase("仲裁归纳：区分一致项与分歧项");
+const arbitrator = agent("仲裁人", {
+  system:
+    "你是仲裁人：把各位专家的表态归纳为『一致项』（所有人都认可，视为稳定结论）与『分歧项』（列明各方论据，不强行折中），" +
+    "并给出最终建议。宁可如实呈现分歧，不要制造虚假共识。用中文回答。",
+});
+const positionsOf = () =>
+  active.map((idx, m) => ({
+    专家: labelOf(idx),
+    position: answers[m].position,
+    keyPoints: answers[m].keyPoints,
+    confidence: answers[m].confidence,
+  }));
+let consensus = await arbitrator.ask<Consensus>(
+  `问题：${question}\n专家表态（匿名）：${JSON.stringify(positionsOf())}\n相互意见：${JSON.stringify(critiques)}\n请归纳共识与分歧。`,
+);
+let rate = computeRate(consensus);
+// 逐轮共识率序列（schema v2：台账记录收敛轨迹）
+const roundsSeq: number[] = [Math.round(rate * 1000) / 1000];
+let round = 1;
+report({ round, rate, target: threshold }, "consensus");
+report({ ratePct: Math.round(rate * 100) + "%", round, alive: `${active.length}/${lenses.length}`, profileLabel: `${profile}` }, "stats");
+log(`第 1 轮共识率 ${Math.round(rate * 100)}%（一致 ${consensus.agreed.length} / 分歧 ${consensus.disputed.length}）`);
+
+phase("修订表态与再仲裁：共识率不足自动加辩");
+while (rate < threshold && round < maxRounds) {
+  round += 1;
+  const revised = await Promise.allSettled(
+    active.map((idx, m) =>
+      experts[idx].ask<Revision>(
+        `仲裁归纳出的分歧项：${JSON.stringify(consensus.disputed)}\n你此前的立场：${answers[m].position}\n` +
+          "请针对分歧重新表态（可坚持，被说服则修改并说明）。",
+      ),
+    ),
+  );
+  const revisedPositions = active.map((idx, m) => {
+    const r = revised[m];
+    if (r.status === "fulfilled") {
+      answers[m] = { position: r.value.position, keyPoints: r.value.keyPoints, confidence: answers[m].confidence };
+      report({ name: lenses[idx], stage: "修订表态", position: r.value.position }, "panel");
+      return { 专家: labelOf(idx), position: r.value.position, keyPoints: r.value.keyPoints, changed: r.value.changed };
+    }
+    return { 专家: labelOf(idx), position: answers[m].position, keyPoints: answers[m].keyPoints, changed: false };
+  });
+  consensus = await arbitrator.ask<Consensus>(
+    `第 ${round} 轮专家表态（匿名）：${JSON.stringify(revisedPositions)}\n上一轮归纳：${JSON.stringify(consensus)}\n请重新归纳一致项与分歧项。`,
+  );
+  rate = computeRate(consensus);
+  roundsSeq.push(Math.round(rate * 1000) / 1000);
+  report({ round, rate, target: threshold }, "consensus");
+  report({ ratePct: Math.round(rate * 100) + "%", round, alive: `${active.length}/${lenses.length}`, profileLabel: `${profile}` }, "stats");
+  log(`第 ${round} 轮收敛后共识率 ${Math.round(rate * 100)}%`);
+}
+active.forEach((idx, m) => report({ name: lenses[idx], stage: "已收敛", position: answers[m].position }, "panel"));
+
+phase("主席综合：产出执行摘要与最终建议");
+const chair = agent("主席", {
+  system:
+    "你是专家团主席。综合专家立场、互评排序与仲裁归纳，为用户写终稿：执行摘要 + 明确的最终建议；" +
+    "分歧未解决时如实呈现分歧而非硬折中；有效专家不足 2 位时须明示结论未经多方独立验证。用中文回答。",
+});
+const rankings = critiques.map((c, m) => ({ 评审: labelOf(active[m]), ranking: c.ranking }));
+const chairOut = await chair.ask<ChairSummary>(
+  `问题：${question}\n有效专家 ${active.length}/${lenses.length}\n` +
+    `专家最终立场（匿名）：${JSON.stringify(positionsOf())}\n互评排序：${JSON.stringify(rankings)}\n` +
+    `最终仲裁归纳：${JSON.stringify(consensus)}\n请给出执行摘要与最终建议。`,
+);
+
+phase("事实复核与落账：复核可验证一致项，记录本次运行");
+const verification = await agent("事实复核员", {
+  system:
+    "你是事实复核员：只复核可客观验证的事实型断言（读代码/查文件/跑命令能判定的就去做；" +
+    "给了证据包则逐条对证并标注出处）；价值判断类结论归入 unverifiable 原样保留。不改任何文件。用中文回答。",
+}).ask<Verification>(`问题语境：${question}\n待复核一致项：${JSON.stringify(consensus.agreed)}${evidenceBlock}${basisBlock}`);
+
+const stable = rate >= threshold;
+// schema v2（2026-10-23 P0 修复）：双口径共识率——rate_pre=仲裁时点（复核前）、rate_post=复核后重算，
+// 消除旧版三时点混用（rate 不可复算）问题。tokens 由回填脚本补齐；adopted 为采纳标记预留。
+const ratePre = Math.round(rate * 1000) / 1000;
+const ratePostDenom = verification.agreed.length + consensus.disputed.length;
+const ratePost = ratePostDenom === 0 ? 0 : Math.round((verification.agreed.length / ratePostDenom) * 1000) / 1000;
+const ledgerEntry = {
+  schema: 2,
+  version: "0.1.2",
+  question: question.slice(0, 120),
+  question_id: String(args.question_id ?? ""),
+  profile,
+  model: "session-default",
+  experts: lenses.length,
+  alive: active.length,
+  rounds: round,
+  rounds_seq: roundsSeq,
+  threshold,
+  rate_pre: ratePre,
+  rate_post: ratePost,
+  rate: ratePre,
+  agreed: verification.agreed.length,
+  disputed: consensus.disputed.length,
+  rejected: verification.rejected.length,
+  unverifiable: verification.unverifiable.length,
+  stable,
+  tokens: null,
+  adopted: null,
+};
+const LEDGER_CODE = [
+  "import sys, json, os, datetime",
+  "p = os.path.expanduser(sys.argv[2])",
+  "d = os.path.dirname(p)",
+  "if d: os.makedirs(d, exist_ok=True)",
+  "e = json.loads(sys.argv[1]); e['ts'] = datetime.datetime.now().isoformat(timespec='seconds')",
+  "open(p, 'a', encoding='utf-8').write(json.dumps(e, ensure_ascii=False) + '\\n')",
+  "print(p)",
+].join("\n");
+const ledger = await world.run("python3", ["-c", LEDGER_CODE, JSON.stringify(ledgerEntry), ledgerPath]);
+log(ledger.exitCode === 0 ? `台账已追加：${ledger.stdout.trim()}` : `台账写入失败（不阻断结论）：${ledger.stderr.slice(0, 120)}`);
+const sampleNote = absent.length > 0
+  ? `样本完整度 ${active.length}/${lenses.length}（缺席：${absent.join("、")}），结论未经全体专家独立验证`
+  : active.length < 2
+    ? "仅 1 位有效专家，共识率反映的是无异议而非多方独立验证"
+    : "";
+
+await artifact.markdown(
+  "summary",
+  [
+    `# 专家团结论·${question.slice(0, 40)}`,
+    "",
+    `- 共识率：${Math.round(rate * 100)}%（阈值 ${Math.round(threshold * 100)}%，${round} 轮收敛，${stable ? "达成共识" : "未达阈值"}）`,
+    `- 档位 ${profile} · 有效专家 ${active.length}/${lenses.length}：${active.map((i) => lenses[i]).join("、")}`,
+    `- 一致项（稳定结论）：${consensus.agreed.length} 条；分歧项：${consensus.disputed.length} 条；复核剔除：${verification.rejected.length} 条`,
+    ...(sampleNote ? [`- ⚠️ ${sampleNote}`] : []),
+    "",
+    "## 一致项（经复核）",
+    ...verification.agreed.map((s) => `- ✅ ${s}`),
+    ...verification.unverifiable.map((s) => `- ❓ ${s}（价值判断/无法客观复核）`),
+    "",
+    "## 仍存分歧（明示，不强行折中）",
+    ...consensus.disputed.map((d) => `- ⚠️ ${d.topic}：${d.sides}`),
+    ...(verification.rejected.length > 0
+      ? ["", "## 复核未通过", ...verification.rejected.map((r) => `- ❌ ${r.item}（${r.reason}）`)]
+      : []),
+    "",
+    "## 执行摘要（主席）",
+    chairOut.executiveSummary,
+    "",
+    "## 最终建议（主席）",
+    chairOut.finalRecommendation,
+    "",
+    "## 仲裁建议（附录）",
+    consensus.recommendation,
+  ].join("\n"),
+  {
+    title: `专家团结论·${question.slice(0, 24)}`,
+    primary: true,
+    description: `共识率 ${Math.round(rate * 100)}%，一致 ${consensus.agreed.length} 条 / 分歧 ${consensus.disputed.length} 条，有效专家 ${active.length}/${lenses.length}。`,
+  },
+);
+
+return {
+  conclusion: stable
+    ? `专家团达成共识（共识率 ${Math.round(rate * 100)}%，${round} 轮）：${verification.agreed.length} 条稳定结论 + ${consensus.disputed.length} 条明示分歧；仲裁建议已给出。${sampleNote ? "⚠️ " + sampleNote : ""}`
+    : `共识率 ${Math.round(rate * 100)}% 低于阈值，未强行合成：${consensus.disputed.length} 条分歧已列明各方论据，建议人工裁决。${sampleNote ? "⚠️ " + sampleNote : ""}`,
+  agreed: verification.agreed,
+  unverifiable: verification.unverifiable,
+  disputed: consensus.disputed,
+  rejected: verification.rejected,
+  executiveSummary: chairOut.executiveSummary,
+  recommendation: chairOut.finalRecommendation || consensus.recommendation,
+  arbiterNote: consensus.recommendation,
+  agreementRate: rate,
+  rounds: round,
+  profile,
+  participants: active.map((i) => lenses[i]),
+  absent,
+  verified: [
+    "每位专家独立作答后才进入互评（无从众）",
+    "互评输入只含匿名标签/立场/要点，不含视角名（防权威偏置）",
+    "共识率由脚本按一致项/(一致项+分歧项) 确定性计算，非模型自报",
+    "可验证事实经事实复核员把关",
+    ...(basis.length > 0 ? ["评测依据已按 basis 参数注入，结论要求逐条对照依据"] : []),
+    ledger.exitCode === 0 ? `运行台账已追加 ${ledger.stdout.trim()}` : "台账写入失败，本次运行未入账",
+  ],
+  notCovered: [
+    stable ? "" : "共识不足，结论未定",
+    "价值判断类结论无法客观复核，仅代表专家团意见",
+    absent.length > 0 ? `缺席专家视角未覆盖：${absent.join("、")}` : "",
+  ].filter(Boolean),
+};
diff --git a/marketplace.json b/marketplace.json
index 350c942..740636d 100644
--- a/marketplace.json
+++ b/marketplace.json
@@ -1386,6 +1386,25 @@
         "unity-cli",
         "skills"
       ]
+    },
+    {
+      "name": "expert-panel",
+      "displayName": "专家团共识评审 Expert Panel",
+      "displayName_i18n": {
+        "en": "Expert Panel Consensus Review",
+        "zh-CN": "专家团共识评审"
+      },
+      "version": "0.1.2",
+      "source": "./external_plugins/expert-panel",
+      "description": "把重要决策交给多位不同视角专家：独立作答→真匿名互评→仲裁归纳（共识率由脚本确定性计算，不足自动加辩）→主席综合→事实复核，产出稳定一致项+明示分歧，绝不硬合成。零配置，默认用会话模型。",
+      "description_i18n": {
+        "en": "Expert-panel consensus review: independent answers from multiple expert lenses → truly anonymous cross-critique → arbitration (consensus rate computed by script, extra debate round below threshold) → chair synthesis → fact verification; quick/standard/deep profiles, zero config, conclusions separate agreed items from explicit disagreements.",
+        "zh-CN": "专家团共识评审：多位不同视角专家独立作答→真匿名互评→仲裁归纳（共识率脚本计算、不足自动加辩）→主席综合→事实复核；quick/standard/deep 三档，零配置，结论按一致/分歧分离绝不硬合成。"
+      },
+      "category": "productivity",
+      "author": {
+        "name": "Zhang24y"
+      }
     }
   ]
 }