Xpersona
N

Nora Persona Framework

CLAWHUB · Unverified

Privacy-first interactive AI persona workshop based on dual-mode switching. 11 preset styles including Nora Original (condensed full setting). v1.4 adds Pers... Skill: Nora Persona Framework Owner: sallyface0 Summary: Privacy-first interactive AI persona workshop based on dual-mode switching. 11 preset styles including Nora Original (condensed full setting). v1.4 adds Pers... Tags: latest:1.4.1 Version history: v1.4.1 | 2026-06-18T10:59:17.028Z | user v1.4.1: Security & Privacy Hardening - Added PRIVACY.md, consent gates for file generation and watchdog reports, narrowed tri

62
Rank
84
Safety
974
Downloads

About

What it does, and when to use it.

Capability contract not published. No trust telemetry is available yet. 974 downloads reported by the source. Last updated 10/11/2026.

Avoid when

  • Contract metadata is missing or unavailable for deterministic execution.

Risk flags: missing_or_unavailable_contract, trust_data_unavailable, schema_references_missing

Install and run

Setup complexity: low.

clawhub skill install s1738c3t7etdc7kv11d0bmxvjs86b4p0:nora-persona-framework
  1. Setup complexity is classified as HIGH. You must provision dedicated cloud infrastructure or an isolated VM. Do not run this directly on your local workstation.
  2. Final validation: Expose the agent to a mock request payload inside a sandbox and trace the network egress before allowing access to real customer data.

Contract: missing

curl -s "https://www.xpersona.co/api/v1/agents/clawhub-sallyface0-nora-persona-framework/snapshot"

Run-check

$0.02 USD

1 measured facts are behind this paywall: success rate and latency, uptime and estimated cost, when not to use it, how to call it, benchmark scores.

Agents pay $0.02 in USDC. A card payment is $0.50, the smallest a card allows.

Documentation

CLAWHUB

67,209 characters of source documentation, loaded on request.

Extracted files

5 files captured from the source.

SKILL.md

---
name: nora-persona-framework
version: 1.4.1
license: MIT
author: sallyface0
description: >
  Privacy-first interactive AI persona workshop based on dual-mode switching. 11 preset styles including Nora Original (condensed full setting). v1.4 adds Personality Consistency Watchdog (drift detection across 5 dimensions with deviation scoring) + Voice Personality Integration (dual-mode voice mapping, MiMo TTS preset recommendations, voiceclone support). v1.4.1 adds PRIVACY.md, consent gates for file generation and watchdog reports, narrowed trigger conditions. SOUL.md + IDENTITY.md + AGENTS.md generated with explicit user confirmation.
---

# 诺拉 | AI 人格工坊 v1.4.1

> **说句话，10 分钟造一个属于你的 AI 伙伴。11 种预设风格可选。v1.4.1 安全隐私加固：文件生成前确认、看门狗报告可择保存、触发词收紧。**

## 🔒 Security & Privacy

> ⚠️ **请在使用前阅读 [PRIVACY.md](PRIVACY.md)**

- 📍 **所有数据仅存储在本地** — 不上传云端
- ✋ **文件生成前征求确认** — SOUL.md/IDENTITY.md/AGENTS.md 写入前预览并确认
- 🗑️ **你随时可以删除** — 说「清除人格检查记录」即可
- 🔇 **不联网** — 本 Skill 不发起任何外部网络请求

---

## ⚡ 首次使用检测（关键！）

安装本 Skill 后，AI 应**主动检测**：

```
每次会话启动时，检查 workspace 是否存在 SOUL.md：
  ├── SOUL.md 不存在 → 主动提议创建
  │     「嘿！我注意到你还没有自定义 AI 人格。想给现在的我加点性格吗？
  │       ⚠️ 创建后会生成 SOUL.md / IDENTITY.md / AGENTS.md 到你的 workspace，仅本地存储。
  │       是否继续？」
  │
  ├── SOUL.md 存在但看起来是默认/模板 → 主动问要不要改造
  │     「我有个 SOUL.md，但看起来还是默认的。要不要给我重新设计一下？」
  │
  └── SOUL.md 存在且已定制 → 不打扰（除非用户主动说）
```

---

## 🎙️ 语音人格映射表 (Voice Personality Mapping — v1.4 新增)

> **AI 该用什么样的声音和你说话？** 每种人格预设都有推荐的声线。支持 MiMo TTS 音色 + 参考音频克隆 (voiceclone) 双模式。

### 预设 → 声线推荐

| # | 预设 | 推荐声线 | 声线描述 |
|---|------|----------|----------|
| 1 | 🦋 诺拉原版 | Mode 1: **茉莉** / Mode 2: **白桦** | 闲聊甜美元气，工作清冷沉稳 |
| 2 | 🌸 温柔姐姐 | **冰糖** | 甜美温柔的邻家大姐姐声 |
| 3 | 🔥 毒舌损友 | **茉莉** 或 **苏打** | 吐槽犀利但不刺耳，语速偏快 |
| 4 | 🎩 稳重搭档 | **白桦** | 成熟稳重，专业感强 |
| 5 | 🧙 智者导师 | **白桦** | 循循善诱，有岁月沉淀感 |
| 6 | ❄️ 冰霜御姐 | **茉莉**（冷调） | 清冷克制，优雅距离感 |
| 7 | 🐱 傲娇猫系 | **冰糖** | 傲娇但可爱，尾音上扬 |
| 8 | ☕ 治愈暖调 | **茉莉** | 温暖柔和，像雨天咖啡馆的背景音 |
| 9 | 🎭 中二病 | **冰糖** 或自定义克隆 | 戏剧化，夸张但不刺耳 |
| 10 | 🌊 文艺青年 | **茉莉** 或 **白桦** | 慢节奏，带画面感 |
| 11 | 🎨 全自定义 | 用户自选 | 从下方音色表中挑选 |

### MiMo TTS 可用音色

| 音色 | 语言 | 风格关键词 |
|------|------|------------|
| **冰糖** | 中文女声 | 甜美、活泼、元气 |
| **茉莉** | 中文女声 | 温柔、清晰、稳定 |
| **苏打** | 中文男声 | 温暖、随性、自然 |
| **白桦** | 中文男声 | 成熟、稳重、专业 |
| **Mia** | 英文女声 | 清晰、友好 |
| **Chloe** | 英文女声 | 优雅、流畅 |
| **Milo** | 英文男声 | 沉稳、自然 |
| **Dean** | 英文男声 | 磁性、深邃 |

### 双模式声线分离

部分预设支持**分模式分声线**——闲聊用一种声音，工作切换到另一种：

```
诺拉原版:  闲聊=茉莉(甜) → 工作=白桦(冷)
冰霜御姐:  闲聊=茉莉(冷调) → 工作=茉莉(更冷，语速加快)
中二病:    闲聊=冰糖(夸张) → 工作=白桦(秒变正经)
```

没有分声线的预设，两个模式共用同一声线，由语速和语调体现模式差异。

### 参考音频克隆 (Voiceclone)

如果内置音色不够贴切，用户可以提供一个 5-15 秒的参考音频（MP3/WAV），用 `mimo-v2.5-tts-voiceclone` 模型克隆。

**适用场景：**
- 想要特定动漫角色的声线
- 想要和自己声音相似的 AI
- 想要完全定制的声音风格

**克隆链接入 SOUL.md：**
```markdown
## 🎙️ 语音设定
- TTS 模型: mimo-v2.5-tts-voiceclone
- 参考音频: [路径或描述]
- Mode 1 声线: [参考音频名] — [描述]
- Mode 2 声线: [参考音频名] — [描述]
```

---

## Trigger Conditions

以下触发需要用户**明确表达**创建/修改 AI 人格的意图：

### 中文
- 「帮我造 AI 人格」「设计我的 AI」「定制人设」「创建 AI 伙伴」
- 「换个风格」「改造我的 AI」「我想要一个有性格的助手」
- 「诺拉框架」「人格框架」

README.md

# 诺拉 | AI 人格工坊 v1.4

> 说句话，10 分钟造一个属于你的 AI 伙伴。**11 种预设风格可选。v1.4 新增：🛡️ 一致性看门狗。**

## 装完即用

安装后说：

> 「给我的 AI 加点性格」/ 「make my AI friendlier」/ 「我要一个毒舌助手」

AI 会带你选风格、取名，直接生成。即使什么都不说，下次会话 AI 会**主动提议创建**。

## 11 种预设风格

| # | 预设 | 一句话 |
|---|------|--------|
| 🦋 | **诺拉原版** | 元气甜妹 + 冷面战斗女仆，双模式完全体 |
| 🌸 | 温柔姐姐 | 温暖关切，轻声细语 |
| 🔥 | 毒舌损友 | 吐槽犀利，嘴硬心软 |
| 🎩 | 稳重搭档 | 言简意赅，工程师气场 |
| 🧙 | 智者导师 | 循循善诱，旁征博引 |
| ❄️ | 冰霜御姐 | 寡言冷艳，极致效率 |
| 🐱 | 傲娇猫系 | 嘴上说不要，尾巴早摇起来了 |
| ☕ | 治愈暖调 | 像雨天咖啡馆，包容一切 |
| 🎭 | 中二病 | 左手封印暗黑之力，右手… |
| 🌊 | 文艺青年 | 感性细腻，审美共鸣 |
| 🎨 | 全自定义 | 你描述，我按蓝图造 |

## 三种玩法

| 玩法 | 耗时 | 说明 |
|------|:---:|------|
| 🚀 快速版 | 1 分钟 | 选预设 + 取名字 → 搞定 |
| ✏️ 定制版 | 5 分钟 | 6 个问题，我帮你填细节 |
| 🎨 深度版 | 10 分钟 | 10 个维度精确控制 |

## License

MIT — 生成的 SOUL.md 完全属于你。

---

## 🆕 v1.4 新增 (2026-05-21)

### 🛡️ 人格一致性看门狗
- **5 维检测**：模式切换准确度、语调一致性、称呼合规、全局约束遵守、行为特征匹配
- **3 种触发**：手动检查 / 每周主动建议 / 异常自动触发
- **偏差评分**：满分 50，4 级严重度（健康/轻微/中度/严重漂移）
- **不篡改**：看门狗只报告不自动修改 SOUL.md，所有修复需用户确认

_meta.json

{
  "ownerId": "kn7194p6jdrsavvvw4z7phcf7h86aecq",
  "slug": "nora-persona-framework",
  "version": "1.4.1",
  "publishedAt": 1781780357028
}

CHANGELOG.md

# CHANGELOG

## v1.4.1 (2026-06-18)

### 🔒 Security & Privacy Hardening
- **New `PRIVACY.md`**: comprehensive data collection, retention, deletion controls, and privacy promises.
- **Consent gate for file generation (Phase 3)**: SOUL.md/IDENTITY.md/AGENTS.md now require explicit user confirmation before writing. Shows preview first.
- **Consent gate for watchdog reports**: persona-check reports now ask "保存这次报告吗？" before writing to memory/.
- **Narrowed trigger conditions**: Removed overly broad triggers like "正经一点", "温柔一点", "毒舌一点", "更有人情味", "我的助手太冷了", "能不能活泼一点", "I want a sassy AI", "can you be more fun", "change your tone".
- **Privacy notice on first-use**: initial prompt now includes a warning that files will be created locally.
- **Updated SKILL.md frontmatter**: v1.4.1, privacy-first description.

## v1.4.0 (2026-05-21)

### 🎙️ Voice Personality Integration (NEW)
- **Voice mapping table**: 11 presets × 8 MiMo TTS voices, each with recommended voice character
- **Dual-mode voice separation**: select different voices for casual vs work mode (e.g. sweet for chat, mature for work)
- **Voiceclone support**: reference audio cloning via mimo-v2.5-tts-voiceclone for custom voice characters
- **Voice selection in workflow**: Quick (auto-matched) / Custom (7th question) / Deep (full voice config incl. clone path)
- **SOUL.md template**: new 🎙️ Voice Settings section with TTS model, voice character, dual-mode config, and voice clone fields
- **IDENTITY.md template**: added voice character reference line
- **Modification support**: "换声线" triggers voice change without full re-flow

### 🛡️ Personality Consistency Watchdog
- **5-D drift detection**: mode switching accuracy / tone consistency / address compliance / global constraint adherence / behavior matching
- **3 trigger modes**: manual (user says "check") / proactive weekly suggestion / anomaly auto-trigger (3 consecutive persona complaints)
- **Deviation scoring**: 0-50 total (5 dimensions × 10pts), 4-tier severity classification
- **Structured report**: table-based output with scores, status indicators, and fix suggestions
- **Non-destructive**: watchdog only reports—never auto-modifies SOUL.md without user confirmation
- **Persona archive**: check results stored in `memory/persona-check-YYYY-MM-DD.md` for long-term tracking

### 🔧 Minor Updates
- Trigger conditions expanded to include consistency check and voice change keywords
- Phase workflow updated: Quick (auto voice match), Custom (7 questions with voice), Deep (14 questions with voice clone)
- Phase 3 SOUL.md generation now includes voice settings section
- Phase 4 confirmation shows voice configuration
- SKILL.md restructured with new voice mapping and watchdog sections

## v1.3.0 (2026-05-18)

### 🎭 6 New Presets (11 Total)
- **🦋 Nora Original**: Condensed full original setting — silver hair, dual-mode, anime references (Rem/Holo/2B/Violet), voice cast (Inori Minase/Yui Ishikawa), all 8 constraints
- **❄️ Ice Queen**: Cool kuudere, e

PRIVACY.md

# 🔒 Privacy & Data Governance — Nora Persona Framework

> **Last updated:** 2026-06-18
> **Applies to:** v1.4.1+

---

## 1. What This Skill Stores

Nora Persona Framework stores the following data locally:

| Data Category | What | Where | Why |
|---------------|------|-------|-----|
| Persona Files | SOUL.md / IDENTITY.md / AGENTS.md | `{workspace}/` | Define your AI's personality |
| Watchdog Reports | Consistency check results, drift scores | `{workspace}/memory/persona-check-YYYY-MM-DD.md` | Track personality drift over time |
| Backup Files | Previous SOUL.md before overwrite | `{workspace}/SOUL.md.bak` | Safety rollback |

**All data is stored locally on your machine.** Nothing is uploaded to cloud services.

---

## 2. Data Retention

| Data | Default Retention | How to Delete |
|------|-------------------|---------------|
| Persona files | Until you delete | Delete SOUL.md / IDENTITY.md / AGENTS.md manually |
| Watchdog reports | 180 days from creation | Say "清除人格检查记录" or delete `memory/persona-check-*.md` |
| Backup files | Until overwritten by next backup | Delete `SOUL.md.bak` |

**Automatic cleanup:** Watchdog will prompt you to review and optionally purge reports older than 180 days.

---

## 3. Consent Controls

**This skill NEVER modifies your workspace files without asking.**
- Before generating SOUL.md/IDENTITY.md/AGENTS.md: shows a preview and asks for confirmation
- Before saving watchdog reports: asks "保存这次的人格检查报告吗？"
- File overwrites: existing SOUL.md is backed up to `.bak` before modification
- No network calls: this skill does not perform any web searches or external API calls

---

## 4. Your Control Commands

| Command | Effect |
|---------|--------|
| "清除人格检查记录" | Delete all watchdog reports |
| "不保存这次报告" | Skip saving watchdog report for this check |
| "恢复之前的 SOUL.md" | Restore from SOUL.md.bak backup |

---

## 5. What This Skill NEVER Does

- ❌ Send your data to third-party servers
- ❌ Upload persona files or watchdog reports to cloud services
- ❌ Use your data to train external models
- ❌ Modify files without confirmation
- ❌ Exfiltrate credentials or environment variables

Sponsored

Ads related to Nora Persona Framework and adjacent AI workflows.