master-skill 0.10.0 → 0.10.1

package/.claude-plugin/marketplace.json

@@ -9,7 +9,7 @@
     {
       "name": "master-skill",
       "description": "FoJin-powered Buddhist AI persona framework — source-grounded, boundary-aware, fidelity-tested, runtime-ready. 15 prebuilt masters across 印度/汉传/藏传/南传 plus compare, debate, and curriculum meta-skills.",
-      "version": "0.10.0",
+      "version": "0.10.1",
       "source": "./",
       "author": {
         "name": "xr843",

package/.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "master-skill",
   "description": "FoJin-powered Buddhist AI persona framework — source-grounded, boundary-aware, fidelity-tested, runtime-ready. 15 prebuilt masters across 印度/汉传/藏传/南传 plus compare, debate, and curriculum meta-skills.",
-  "version": "0.10.0",
+  "version": "0.10.1",
   "author": {
     "name": "xr843",
     "email": "xr843@users.noreply.github.com"

package/.cursor-plugin/plugin.json

@@ -2,7 +2,7 @@
   "name": "master-skill",
   "displayName": "Master Skill",
   "description": "FoJin-powered Buddhist AI persona framework — source-grounded, boundary-aware, fidelity-tested, runtime-ready. 15 prebuilt masters across 印度/汉传/藏传/南传.",
-  "version": "0.10.0",
+  "version": "0.10.1",
   "author": {
     "name": "xr843",
     "email": "xr843@users.noreply.github.com"

package/ETHICS.md

@@ -8,12 +8,12 @@
 
 **所有通过 Master-skill 生成的对话、文本、回答，均为 AI 合成内容，不是真实祖师的著作、开示或教言。**
 
-- 每位预置法师的回答均由大型语言模型基于 CBETA 经典文献 + `teaching.md` / `voice.md` 合成，**不代表**历史上该法师的原话、亲口开示或亲笔著作
+- 每位预置法师的回答均由大型语言模型基于该 persona 声明的原典来源（CBETA、BDRC / Toh、SuttaCentral / PTS 或合规编纂开示）+ `teaching.md` / `voice.md` 合成，**不代表**历史上该法师的原话、亲口开示或亲笔著作
 - AI 角色对祖师表达风格的还原是**近似**，非权威：语言选词、句式节奏、比喻用法由模型生成，不可直接引用为"某法师说过"
-- 所有引经据典的 CBETA 编号来自 `sources/` 离线片段或 FoJin 实时检索，但**回答中的文义阐释**是 AI 组合生成，可能与祖师原文含义有偏差
+- 所有引经据典的来源标识必须来自 persona 的 `meta.json.sources[]`；即使来源标识真实，**回答中的文义阐释**仍是 AI 组合生成，可能与祖师原文含义有偏差
 - 使用时请始终默认："这是基于文献的 AI 学习辅助"，不是"与祖师对话"。前者是工具，后者是误解
 
-如你在任何公开场合引用、转发、发表由本项目生成的文本，**必须明确标注 AI 生成属性与原始出处**（CBETA 经号 / FoJin 链接）。将 AI 生成内容作为祖师原话传播，既违反本协议，也违背佛教"不妄语"的基本戒律。
+如你在任何公开场合引用、转发、发表由本项目生成的文本，**必须明确标注 AI 生成属性与原始出处**（来源家族、声明 ID、题名及可用定位信息）。将 AI 生成内容作为祖师原话传播，既违反本协议，也违背佛教"不妄语"的基本戒律。
 
 ---
 
@@ -171,7 +171,7 @@ Master-skill 采取**双轨授权**，代码与内容分开授权：
 | 源代码（`scripts/`、`tools/`、`bin/`、`hooks/`、`.github/`、workflow、CI） | **MIT** | 任意使用、修改、商用 | 去除版权声明 |
 | 预置法师内容（`prebuilt/**/SKILL.md`、`teaching.md`、`voice.md`、`sources/*.md`、`fidelity.jsonl`） | **CC BY-NC-SA 4.0** | 署名 + 非商用 + 相同方式共享下任意使用 | 未署名、纯商业闭源分发 |
 | Prompts 模板（`prompts/**`） | **CC BY 4.0** | 署名后任意使用 | 去署名 |
-| FoJin 检索返回的原始经文 | **CBETA 知识共用 非商业性 禁止改作 3.0** | 遵循 CBETA 原协议 | 违反 CBETA 协议 |
+| FoJin 检索返回的原始经文 / 元数据 / 编纂开示 | **按返回来源族、具体版本与权利人的原始许可**（CBETA、BDRC / Toh、SuttaCentral / PTS、compiled teachings 各自适用） | 在对应许可与合理引用范围内使用并保留出处 | 把某一来源许可套用于其他来源、移除归属、超授权复制或商用 |
 
 **商业化使用（含但不限于）：**
 - SaaS 付费问答服务嵌入 Master-skill 法师内容
@@ -184,24 +184,30 @@ Master-skill 采取**双轨授权**，代码与内容分开授权：
 
 ## 5. 数据来源透明 (Data Provenance)
 
-每位法师 `SKILL.md` frontmatter 必须声明：
-
-```yaml
-sources:
-  - title: {经典名称}
-    cbeta_id: {CBETA 编号，如 T30n1579}
-    fojin_text_id: {FoJin 内部 ID}
-citation_format: "【《{title}》卷{juan}，{cbeta_id}】"
-verified_by: {维护者 GitHub 用户名}
-verified_at: {YYYY-MM-DD}
+每位 persona 的 `meta.json` 必须声明来源清单与同源 citation contract：
+
+```json
+{
+  "sources": [
+    {"type": "cbeta | tibetan_canon | pali_canon | compiled_teaching | ...", "id": "来源族内的声明 ID", "title": "经典或开示题名"}
+  ],
+  "citation_contract": {
+    "version": 1,
+    "claim_policy": "declared_sources_only",
+    "required_for": ["doctrinal_claim", "practice_guidance", "text_interpretation"],
+    "allowed_source_types": ["由 sources[].type 自动派生"],
+    "minimum_claim_coverage": 0.9,
+    "live_retrieval_allowed": true
+  }
+}
 ```
 
 **HARD-GATE 铁律：**
-- 无 CBETA 经号的教义断言不得写入 `teaching.md`
-- 不得捏造 CBETA 编号（CI `scripts/validate.py` 会对照 FoJin 反查）
+- 教义断言、修行指导与文本解释若不能解析到 `meta.json.sources[]` 的声明来源，不得写入 `teaching.md`
+- 不得捏造来源 ID，来源类型必须属于 `citation_contract.allowed_source_types`
 - 不得为虚构人物、合成 persona、无史实记载者建立 `prebuilt/`
 
-违反以上任一，PR 将被自动驳回。
+`scripts/validate-citation-contract.py` 与 `verify_sources.py --check-links/--final-check` 会离线检查来源清单、家族 ID 与合同一致性；它们不解析正文自由文本，也不保证外部站点可达。违反以上任一，PR 将被自动驳回。
 
 ---
 

package/README.md

@@ -143,10 +143,10 @@ Master-skill 是由 [FoJin](https://fojin.app) 驱动的佛教 AI 祖师人格
 ## 特性
 
 - **预置十五位四大传统祖师**：1 位印度（龙树 · 中观）+ 8 位汉传（唯识、中观、禅、天台、华严、净土、跨宗派）+ 3 位藏传（阿底峡 · 噶当；宗喀巴 · 格鲁；米拉日巴 · 噶举）+ 3 位南传（觉音 · 上座部论师；马哈希 · 缅甸内观；阿姜查 · 泰国森林）—— 另含 `compare-masters` 多祖师对比 meta-skill，开箱即用
-- **经文溯源（Provenance）**：每位祖师附原典标识（CBETA / BDRC / SuttaCentral）+ FoJin text ID，所有教义断言强制附原典引证
+- **经文溯源（Provenance）**：每位祖师附声明来源 ID（CBETA / BDRC / Toh / SuttaCentral / PTS / 合规编纂开示）；实时检索仅在返回真实 `text_id` 时附 FoJin 定位链接，所有教义断言强制附原典引证
 - **离线经文片段**：`sources/` 目录收录核心经典关键段落，FoJin 不可用时仍可离线引用
 - **渐进式披露**：SKILL.md 以决策树 + Quick Ref 为主，`references/`、`sources/` 按需加载，Context 随查随取
-- **HARD-GATE 铁律**：`/create-master` 与预置法师内置红线——无 CBETA 引证的教义断言不得写入、不得捏造经号、不得为虚构人物建角色
+- **HARD-GATE 铁律**：`/create-master` 与预置法师内置红线——教义断言、修行指导与文本解释必须引用该 persona 声明的来源（CBETA / BDRC / Toh / SuttaCentral / PTS / 合规编纂开示），不得捏造来源 ID，不得为虚构人物建角色
 - **二阶段独立审查**：生成管线在写入前强制经过"教义准确性 → 风格一致性"两轮独立审查，FAIL 自动修复最多 2 轮
 - **自动化保真度测试**：每位祖师 `tests/fidelity.jsonl` 10+ 条 Q&A（`compare-masters` 元技能 18 条），验证引用和关键词覆盖；CI 在每次推送时 dry-run 验证（结构校验）；实跑评分需 `ANTHROPIC_API_KEY`，作为本地/发版前手动步骤执行
 - **多平台统一插件**：Claude Code、Cursor、Codex CLI、OpenCode、Gemini CLI 共用一份 `prebuilt/`，session-start hook 跨平台注入法师列表
@@ -179,14 +179,20 @@ Master-skill 的核心不是"角色扮演提示词集合"，而是一个可验
 
 **NPX 一键安装（推荐，无需常驻）**
 
+`npx master-skill install --all` 一次安装全部 19 个 Skill：15 位祖师、3 个教学模式，以及 `create-master` 生成器。`create-master` 会复制自包含的运行时，因此临时 npx 包目录被清理后仍可使用；重新安装或 `update --all` 会更新运行时，但保留 `create-master/masters/` 中用户生成的 persona。
+
 ```bash
 # 安装指定祖师
 npx master-skill install master-zhiyi master-fazang master-huineng
 
-# 安装全部 15 位（四大传统）
+# 单独安装公共教学模式或自定义生成器
+npx master-skill install compare-masters
+npx master-skill install create-master
+
+# 安装全部 19 个 Skill
 npx master-skill install --all
 
-# 查看可用祖师
+# 查看全部可安装 Skill
 npx master-skill list
 ```
 
@@ -205,7 +211,8 @@ npm update -g master-skill             # 升到下一个 minor / patch
 # npx（上方）与 git clone 手动安装为正式发布渠道：
 git clone https://github.com/xr843/Master-skill ~/Master-skill
 cd ~/Master-skill && pip install -r requirements.txt
-for d in prebuilt/*/; do ln -sf "$(pwd)/$d" ~/.claude/skills/"$(basename $d)"; done
+for d in prebuilt/master-*/; do ln -sf "$(pwd)/$d" ~/.claude/skills/"$(basename $d)"; done
+ln -sf "$(pwd)/prebuilt/compare" ~/.claude/skills/compare-masters
 ln -sf "$(pwd)" ~/.claude/skills/create-master
 ```
 
@@ -438,7 +445,7 @@ SKILL.md (AgentSkills 入口：决策树 + Quick Ref)
     |                   |   +-- teaching.md
     |                   |   +-- voice.md
     |                   +-- sources/          (离线经文片段)
-    |                   |   +-- *.md (CBETA 段落)
+    |                   |   +-- *.md (声明来源的离线段落)
     |                   +-- tests/
     |                       +-- fidelity.jsonl  (保真度样例, CI dry-run)
     |
@@ -479,7 +486,7 @@ Master-skill 通过 `tools/fojin_bridge.py` 接入 FoJin API，实现：
 - 语义向量相似度搜索（教义相关经文）
 - 原文段落提取与出处追踪
 
-所有引用均附带可追溯的 FoJin 链接，确保内容来源透明。
+所有引用都必须能追溯到 persona 声明的来源 ID；仅当实时检索返回真实 `text_id` 时附 FoJin 定位链接，否则使用对应官方目录或离线声明来源。
 
 ---
 
@@ -493,8 +500,8 @@ Master-skill 通过 `tools/fojin_bridge.py` 接入 FoJin API，实现：
 
 **要做：**
 
-- 忠实依据经文原文，所有回答附 FoJin 出处链接
-- 通过运行时 RAG 检索真实经文，而非仅依赖 AI 自身知识
+- 忠实依据声明来源，回答附可追溯的来源 ID；有真实 `text_id` 时再附 FoJin 定位链接
+- 仅在 citation contract 允许且离线资料不足时通过运行时 RAG 检索，不以 AI 自身知识冒充原典
 - 遇到超出范围的问题坦诚说明
 
 ---
@@ -505,9 +512,9 @@ Master-skill 通过 `tools/fojin_bridge.py` 接入 FoJin API，实现：
 
 能。每位预置法师的 `prebuilt/<name>/sources/` 收录了该法师核心经典的关键段落（离线经文片段）。FoJin 不可用时，法师会降级到离线模式并在回答中声明"当前使用离线片段"。`/create-master` 管线遇到 API 故障会提示用户切换手动输入模式，由用户粘贴经文原文继续生成。
 
-**Q：CBETA 引用格式是什么样的？怎么验证？**
+**Q：CBETA 引用格式是什么样的？来源如何验证？**
 
-所有 CBETA 引证必须带 `Txxn####` 形式的经号（例如《妙法蓮華經》→ `T9n262`）。`scripts/validate.py` 会检查 frontmatter 的 `sources` 字段格式；`tools/verify_sources.py` 在写入前会逐条核对 FoJin `text_id` 的有效性，失效链接自动降级为 FoJin 搜索链接，不会留下死链。
+CBETA 引证使用 `Txxn####` 形式的经号（例如《妙法蓮華經》→ `T09n0262`）；藏传、南传与编纂开示分别使用 persona 在 `meta.json.sources[]` 中声明的 BDRC / Toh、SuttaCentral / PTS 或 teaching ID。`scripts/validate-citation-contract.py` 与 `tools/verify_sources.py --check-links/--final-check` 离线检查来源家族、ID 格式、声明归属和合同一致性；它们不解析正文自由文本，也不保证外部链接 HTTP 可达。旧版 `verify_sources.py --fix` 在线审计只覆盖 CBETA / FoJin 链接。
 
 **Q：`npx master-skill install` 执行失败、报 ENOTEMPTY 或权限错误怎么办？**
 

package/README_EN.md

@@ -141,10 +141,10 @@ This project is built out of respect for Buddhist traditions. All content is gen
 ## Features
 
 - **15 pre-built masters across four traditions**: 1 印度 (Madhyamaka · Nāgārjuna) + 8 汉传 (Yogācāra, Madhyamaka, Chan, Tiantai, Huayan, Pure Land, cross-tradition) + 3 藏传 (Kadam · Atiśa; Gelug · Tsongkhapa; Kagyu · Milarepa) + 3 南传 (Theravāda commentator · Buddhaghosa; Burmese vipassanā · Mahasi Sayadaw; Thai Forest · Ajahn Chah) — plus a `compare-masters` multi-master comparison meta-skill, ready to use out of the box
-- **Provenance enforcement**: Every master ships with authoritative source IDs (CBETA / BDRC / SuttaCentral) and FoJin text IDs in frontmatter; every doctrinal claim must carry a scriptural citation
+- **Provenance enforcement**: Every master ships with declared source IDs (CBETA / BDRC / Toh / SuttaCentral / PTS / compliant compiled teachings); live retrieval adds a FoJin locator only when a real `text_id` is returned, and every doctrinal claim must carry a source citation
 - **Offline source passages**: `sources/` captures key passages from each master's core canon, so citations still work when FoJin is unreachable
 - **Progressive disclosure**: SKILL.md is a decision tree + quick reference; `references/` and `sources/` are loaded on demand to keep context lean
-- **HARD-GATE discipline**: Both `/create-master` and every prebuilt master embed hard rules — no unverified CBETA ID, no uncited doctrinal claim, no fictional personas
+- **HARD-GATE discipline**: Both `/create-master` and every prebuilt master require doctrinal claims, practice guidance, and text interpretation to cite that persona's declared sources (CBETA / BDRC / Toh / SuttaCentral / PTS / compliant compiled teachings); fabricated source IDs and fictional personas are forbidden
 - **Two-stage independent review**: The generation pipeline forces a "doctrinal accuracy → voice consistency" review before write; FAIL triggers up to 2 rounds of automatic repair
 - **Automated fidelity tests**: Each master's `tests/fidelity.jsonl` holds 10+ Q&A samples (the `compare-masters` meta-skill holds 18) validating citations and keyword coverage; CI runs a dry-run on every push
 - **Unified multi-platform plugin**: Claude Code, Cursor, Codex CLI, OpenCode, and Gemini CLI share one `prebuilt/` tree, with a session-start hook injecting the master list on every platform
@@ -177,9 +177,17 @@ The v1.0 track prioritizes framework stability over adding more masters. See [do
 
 **NPX (recommended, no global state)**
 
+`npx master-skill install --all` installs all 19 skills: 15 personas, 3 teaching modes, and the `create-master` generator. The generator is copied as a self-contained runtime, so it remains usable after the transient npx package directory is removed; reinstall and `update --all` refresh the runtime while preserving user-generated personas under `create-master/masters/`.
+
 ```bash
-npx master-skill install --all    # Install all 15 masters
-npx master-skill list             # List available masters
+# Install individual public skills
+npx master-skill install master-zhiyi
+npx master-skill install compare-masters
+npx master-skill install create-master
+
+# Install or list the complete 19-skill catalog
+npx master-skill install --all
+npx master-skill list
 ```
 
 **Global install (frequent use / offline-friendly)**
@@ -196,7 +204,8 @@ npm update -g master-skill             # Pull next minor / patch
 ```bash
 git clone https://github.com/xr843/Master-skill ~/Master-skill
 cd ~/Master-skill && pip install -r requirements.txt
-for d in prebuilt/*/; do ln -sf "$(pwd)/$d" ~/.claude/skills/"$(basename $d)"; done
+for d in prebuilt/master-*/; do ln -sf "$(pwd)/$d" ~/.claude/skills/"$(basename $d)"; done
+ln -sf "$(pwd)/prebuilt/compare" ~/.claude/skills/compare-masters
 ln -sf "$(pwd)" ~/.claude/skills/create-master
 ```
 
@@ -414,7 +423,7 @@ SKILL.md (AgentSkills entry: decision tree + quick reference)
     |                           +-- references/       (loaded on demand)
     |                           |   +-- teaching.md
     |                           |   +-- voice.md
-    |                           +-- sources/          (offline CBETA passages)
+    |                           +-- sources/          (offline declared-source passages)
     |                           |   +-- *.md
     |                           +-- tests/
     |                               +-- fidelity.jsonl  (CI dry-run samples)
@@ -457,7 +466,7 @@ Master-skill connects to the FoJin API via `tools/fojin_bridge.py` to enable:
 - Runtime RAG retrieval for grounding answers in real texts
 - Source passage extraction with provenance tracking
 
-All citations include traceable FoJin links to ensure transparency of sources.
+Every citation must resolve to the persona's declared source ID. A FoJin locator is added only when live retrieval returns a real `text_id`; otherwise the corresponding official catalog or offline declared source is used.
 
 ---
 
@@ -473,8 +482,8 @@ All citations include traceable FoJin links to ensure transparency of sources.
 
 **Will:**
 
-- Cite source texts faithfully, with FoJin links on every response
-- Retrieve real texts via runtime RAG, not relying solely on AI training data
+- Cite declared sources faithfully with traceable source IDs, adding a FoJin locator only when a real `text_id` is available
+- Use runtime RAG only when the citation contract permits it and offline material is insufficient; never present model memory as a primary text
 - Acknowledge clearly when a question falls outside scope
 - Encourage users to seek out qualified masters and authentic practice
 
@@ -486,9 +495,9 @@ All citations include traceable FoJin links to ensure transparency of sources.
 
 Yes. Each prebuilt master ships with `prebuilt/<name>/sources/` — key passages from that master's core canon, stored offline. When FoJin is down, the master degrades to offline mode and declares "currently running on offline passages" in the reply. The `/create-master` pipeline asks the user to switch to manual-input mode when the API fails, so you can paste source text and continue.
 
-**Q: What does a valid CBETA citation look like, and how is it verified?**
+**Q: What does a valid CBETA citation look like, and how are sources verified?**
 
-Every CBETA citation must carry a `Txxn####` identifier (for example, the Lotus Sutra is `T9n262`). `scripts/validate.py` lints the frontmatter `sources` block; `tools/verify_sources.py` checks every FoJin `text_id` against the live API before writing. Broken links are downgraded to FoJin search URLs — no dead references make it into the final file.
+CBETA citations use a `Txxn####` identifier (for example, the Lotus Sutra is `T09n0262`); Tibetan, Pali, and compiled-teaching personas use the BDRC / Toh, SuttaCentral / PTS, or teaching IDs declared in `meta.json.sources[]`. `scripts/validate-citation-contract.py` and `tools/verify_sources.py --check-links/--final-check` validate source families, identifier shapes, declared membership, and contract consistency offline. They do not parse free-text citations or guarantee HTTP reachability. The legacy online `verify_sources.py --fix` audit covers CBETA / FoJin links only.
 
 **Q: `npx master-skill install` fails with ENOTEMPTY or a permission error — what now?**
 
@@ -508,9 +517,9 @@ See "Contributing" below. The short version: follow the v0.3 layout under `prebu
 
 Contributions are welcome: new prebuilt masters, corrections to source attributions, offline passage additions, or toolchain improvements.
 
-New masters must follow the v0.3 layout: `prebuilt/<name>/` containing SKILL.md (with provenance frontmatter and a decision tree), `references/teaching.md` and `references/voice.md` (loaded on demand), `sources/*.md` (offline CBETA passages), and `tests/fidelity.jsonl` (5+ Q&A fidelity samples). Run `python3 scripts/validate.py --strict` for zero errors, and make sure the CI fidelity dry-run passes before opening a PR.
+New masters must follow the v0.3 layout: `prebuilt/<name>/` containing SKILL.md (with provenance routing and a decision tree), `meta.json` (declared sources plus citation contract), `references/teaching.md` and `references/voice.md` (loaded on demand), `sources/*.md` (offline declared-source passages), and `tests/fidelity.jsonl` (5+ Q&A fidelity samples). Run `python3 scripts/validate.py --strict` and `python3 scripts/validate-citation-contract.py` for zero errors, and make sure the CI fidelity dry-run passes before opening a PR.
 
-Before submitting, verify that sources trace back to CBETA, content is faithful to historical documents, and no sectarian bias is introduced.
+Before submitting, verify that every source resolves to the persona's declared source family, content is faithful to historical documents, and no sectarian bias is introduced.
 
 ---
 

package/SKILL.md

@@ -72,7 +72,7 @@ allowed-tools:
 
 ### Step 2：数据采集
 
-使用 `${CLAUDE_SKILL_DIR}/tools/sutra_collector.py --name "<法师名>" --tradition "<传承>"` 从 FoJin 采集知识图谱实体、经典内容、传承术语。采集后用 `verify_sources.py --check-links` 验证 CBETA / BDRC / SC ID。
+使用 `${CLAUDE_SKILL_DIR}/tools/sutra_collector.py --name "<法师名>" --tradition "<传承>" --output collected_data.json` 从 FoJin 采集知识图谱实体、经典内容、传承术语。采集后运行 `${CLAUDE_SKILL_DIR}/tools/verify_sources.py --check-links collected_data.json`，离线验证来源家族、ID 格式、声明归属与自动派生的 citation contract。
 
 API 故障 / 超时 / 数据阈值 / 引用规则细节 → `references/workflow-details.md` §Step 2 + `references/source-conventions.md`。
 
@@ -84,7 +84,7 @@ API 故障 / 超时 / 数据阈值 / 引用规则细节 → `references/workflow
 
 ### Step 3.5：二阶段审查
 
-教义准确性（`doctrine_reviewer.md`，CBETA 经证覆盖率 ≥ 90%） → 风格一致性（`voice_reviewer.md`，Layer 0 硬规则完整）。审查顺序不可颠倒。FAIL → 自动修复重审，最多 2 轮，仍 FAIL → 人工介入。
+生成器先从 `sources[].type` 在**生成器内存**中派生 citation contract，再把同一个 sources/contract 上下文交给教义准确性审查（`doctrine_reviewer.md`，按 `citation_contract.minimum_claim_coverage` 审核声明来源覆盖率） → 风格一致性（`voice_reviewer.md`，Layer 0 硬规则完整）。审查顺序不可颠倒。FAIL → 自动修复重审，最多 2 轮，仍 FAIL → 人工介入；最终写入的 `meta.json` 必须复用并再次校验同一 contract。
 
 ### Step 4：预览与确认
 
@@ -92,7 +92,7 @@ API 故障 / 超时 / 数据阈值 / 引用规则细节 → `references/workflow
 
 ### Step 5：写入文件
 
-`master_builder.py --name "<法师名>" --output masters/` 写入 `masters/{slug}/{SKILL.md,teaching.md,voice.md,meta.json}`。写入前 `verify_sources.py --final-check` 最终验证，无效链接降级为 FoJin 搜索链接。
+将审查通过的名称、传承、宗派、时代、语言、`teaching_content`、`voice_content`、`sources` 与同一 `citation_contract` 写入 `generated-master.json`，再运行 `${CLAUDE_SKILL_DIR}/tools/master_builder.py --spec generated-master.json --output "${CLAUDE_SKILL_DIR}/masters/"`。生成器统一写入 `masters/master-{slug}/`，且 `SKILL.md` 的 name 为 `master-{slug}`；随后运行 `${CLAUDE_SKILL_DIR}/tools/verify_sources.py --final-check "${CLAUDE_SKILL_DIR}/masters/master-{slug}/"`。该终验离线检查四个必需文件、目录/name 一致性，以及 `meta.json` 的来源清单、家族 ID、声明归属和 contract；它不解析 `teaching.md` 自由文本，也不保证外部站点可达。
 
 OpenClaw / Claude Code 注册路径 → `references/workflow-details.md` §Step 5。
 
@@ -133,8 +133,8 @@ KG 深度遍历 / 跨词典对比等 `rag_query.py` 不够用的场景 → `refe
 
 ## 铁律（HARD-GATE）
 
-- **NO DOCTRINAL CLAIM WITHOUT CBETA CITATION** — teaching.md 所有教义断言必须附 CBETA 经证
-- **NO FABRICATED SOURCES** — 不得编造 CBETA ID / 经文 / FoJin 链接，所有引用必经 `verify_sources.py` 验证
+- **NO DOCTRINAL CLAIM WITHOUT A DECLARED SOURCE CITATION.** — 所有教义断言、修行指导与文本解释的引用必须解析到所选 persona 的 `meta.json.sources[]`，来源类型必须列于 `citation_contract.allowed_source_types`；仅当 `citation_contract.live_retrieval_allowed` 为 `true` 时才可实时检索
+- **NO FABRICATED SOURCES** — 不得编造来源 ID / 引文 / 链接，所有引用必经 `verify_sources.py` 验证
 - **NO FICTIONAL PERSONAS** — 仅历史真实人物，不为虚构角色创建
 
 完整理性化防御表、红旗清单、ETHICS.md 运行时摘要 → `references/ethics-runtime.md`。