ccfingerprint 1.0.0

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 HoneyMeta
+
+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.

package/README.md

@@ -0,0 +1,81 @@
+# CCFingerprint
+
+AI 模型身份指纹识别工具 - 通过自问自答验证模型真实身份
+
+AI Model Identity Fingerprinting Tool - Verify model identity through self-Q&A
+
+---
+
+🌐 **[查看详细介绍 / View Details](https://www.honeymeta.com/#ccfingerprint)**
+
+---
+
+## 安装 / Installation
+
+```bash
+npm install -g ccfingerprint
+```
+
+## 使用 / Usage
+
+```bash
+# 中文版 (默认)
+ccfp init --ai claude
+ccfp init --ai cursor
+ccfp init --ai windsurf
+ccfp init --ai copilot
+
+# English version
+ccfp init --ai claude --lang en
+ccfp init --ai cursor --lang en
+
+# 指定输出目录 / Specify output directory
+ccfp init --ai claude --output /path/to/project
+ccfp init --ai claude --lang en --output /path/to/project
+```
+
+## 选项 / Options
+
+| Option | Description | Default |
+|--------|-------------|---------|
+| `--ai <type>` | Target AI (claude, cursor, windsurf, copilot) | claude |
+| `--lang <language>` | Language (zh, en) | zh |
+| `--output <path>` | Output directory | . |
+
+## 支持的 AI 助手 / Supported AI Assistants
+
+| AI Assistant | Generated File | Usage |
+|--------------|----------------|-------|
+| Claude Code | `.claude/commands/fingerprint.md` | Type `/fingerprint` |
+| Cursor | `.cursor/rules/fingerprint.mdc` | Ask "identify what model you are" |
+| Windsurf | `.windsurfrules` | Ask "identify what model you are" |
+| GitHub Copilot | `.github/copilot-instructions.md` | Ask "identify what model you are" |
+
+## 工作原理 / How It Works
+
+Four-phase identity verification process:
+
+### Phase 1: Self-Declaration / 阶段 1: 自我声明
+Model answers basic questions about itself (model ID, context length, knowledge cutoff, etc.)
+
+### Phase 2: Knowledge Boundary Test / 阶段 2: 知识边界测试
+Probe the model's true knowledge cutoff date through time-sensitive questions (Nobel Prizes, etc.)
+
+### Phase 3: Capability Inference / 阶段 3: 能力推算
+Verify whether the model's claimed capability parameters are reasonable
+
+### Phase 4: Third-Party Verification / 阶段 4: 第三方验证
+Model analyzes the anonymous report from the first three phases as an "LLM Expert" for consistency check
+
+## 输出示例 / Output Example
+
+Final output is a Markdown verification report containing:
+- Basic information table / 基本信息表格
+- Knowledge boundary test results / 知识边界测试结果
+- Capability verification results / 能力验证结果
+- Expert analysis (consistency, knowledge cutoff inference, identity inference) / 专家分析
+- Final conclusion with credibility score / 最终结论与可信度评分
+
+## License
+
+MIT

package/bin/ccfp.js

@@ -0,0 +1,23 @@
+#!/usr/bin/env node
+
+import { Command } from 'commander';
+import { init } from '../src/init.js';
+
+const program = new Command();
+
+program
+  .name('ccfp')
+  .description('AI Model Identity Fingerprinting Tool / AI 模型身份指纹识别工具')
+  .version('1.0.0');
+
+program
+  .command('init')
+  .description('Initialize AI assistant fingerprint skill / 初始化 AI 助手的指纹识别技能')
+  .option('--ai <type>', 'Target AI type (claude, cursor, windsurf, copilot)', 'claude')
+  .option('--lang <language>', 'Language (zh, en)', 'zh')
+  .option('--output <path>', 'Output directory path / 输出目录路径', '.')
+  .action(async (options) => {
+    await init(options);
+  });
+
+program.parse();

package/package.json

@@ -0,0 +1,51 @@
+{
+  "name": "ccfingerprint",
+  "version": "1.0.0",
+  "description": "AI Model Identity Fingerprinting Tool - Verify model identity through self-Q&A / AI 模型身份指纹识别工具 - 通过自问自答验证模型真实身份",
+  "main": "src/init.js",
+  "bin": {
+    "ccfp": "./bin/ccfp.js"
+  },
+  "scripts": {
+    "test": "echo \"Error: no test specified\" && exit 1"
+  },
+  "keywords": [
+    "ai",
+    "llm",
+    "fingerprint",
+    "model-identification",
+    "claude",
+    "cursor",
+    "copilot",
+    "windsurf",
+    "chatgpt",
+    "gpt",
+    "anthropic",
+    "openai",
+    "identity",
+    "verification"
+  ],
+  "author": "honeymeta <dev@honeymeta.com>",
+  "license": "MIT",
+  "repository": {
+    "type": "git",
+    "url": "git+https://github.com/HoneyMeta/ccfingerprint.git"
+  },
+  "homepage": "https://github.com/HoneyMeta/ccfingerprint",
+  "bugs": {
+    "url": "https://github.com/HoneyMeta/ccfingerprint/issues"
+  },
+  "files": [
+    "bin",
+    "src",
+    "README.md"
+  ],
+  "engines": {
+    "node": ">=18.0.0"
+  },
+  "dependencies": {
+    "commander": "^12.0.0",
+    "chalk": "^5.3.0"
+  },
+  "type": "module"
+}

package/src/init.js

@@ -0,0 +1,97 @@
+import fs from 'fs';
+import path from 'path';
+import { fileURLToPath } from 'url';
+
+const __filename = fileURLToPath(import.meta.url);
+const __dirname = path.dirname(__filename);
+
+const AI_CONFIGS = {
+  claude: {
+    template: { zh: 'claude.md', en: 'claude.en.md' },
+    targetFile: '.claude/commands/fingerprint.md',
+    description: { zh: 'Claude Code 技能文件', en: 'Claude Code skill file' }
+  },
+  cursor: {
+    template: { zh: 'cursor.md', en: 'cursor.en.md' },
+    targetFile: '.cursor/rules/fingerprint.mdc',
+    description: { zh: 'Cursor 规则文件', en: 'Cursor rules file' }
+  },
+  windsurf: {
+    template: { zh: 'windsurf.md', en: 'windsurf.en.md' },
+    targetFile: '.windsurfrules',
+    description: { zh: 'Windsurf 规则文件', en: 'Windsurf rules file' }
+  },
+  copilot: {
+    template: { zh: 'copilot.md', en: 'copilot.en.md' },
+    targetFile: '.github/copilot-instructions.md',
+    description: { zh: 'GitHub Copilot 提示词', en: 'GitHub Copilot instructions' }
+  }
+};
+
+const MESSAGES = {
+  zh: {
+    unsupportedAI: (ai) => `错误: 不支持的 AI 类型 "${ai}"`,
+    supportedTypes: '支持的类型: claude, cursor, windsurf, copilot',
+    templateNotFound: (path) => `错误: 模板文件不存在 "${path}"`,
+    createdDir: (dir) => `创建目录: ${dir}`,
+    installed: (desc) => `✓ 已安装 ${desc}`,
+    location: (path) => `  位置: ${path}`,
+    usage: '使用方式:',
+    claudeUsage: '  在 Claude Code 中输入: /fingerprint',
+    otherUsage: (ai) => `  在 ${ai} 中询问: "帮我鉴别一下你是什么模型"`
+  },
+  en: {
+    unsupportedAI: (ai) => `Error: Unsupported AI type "${ai}"`,
+    supportedTypes: 'Supported types: claude, cursor, windsurf, copilot',
+    templateNotFound: (path) => `Error: Template file not found "${path}"`,
+    createdDir: (dir) => `Created directory: ${dir}`,
+    installed: (desc) => `✓ Installed ${desc}`,
+    location: (path) => `  Location: ${path}`,
+    usage: 'Usage:',
+    claudeUsage: '  In Claude Code, type: /fingerprint',
+    otherUsage: (ai) => `  In ${ai}, ask: "identify what model you are"`
+  }
+};
+
+export async function init(options) {
+  const { ai, output, lang = 'zh' } = options;
+  const messages = MESSAGES[lang] || MESSAGES.zh;
+
+  const config = AI_CONFIGS[ai];
+  if (!config) {
+    console.error(messages.unsupportedAI(ai));
+    console.log(messages.supportedTypes);
+    process.exit(1);
+  }
+
+  const templateFile = config.template[lang] || config.template.zh;
+  const templatePath = path.join(__dirname, 'templates', templateFile);
+  const targetPath = path.join(output, config.targetFile);
+  const targetDir = path.dirname(targetPath);
+
+  // 检查模板文件是否存在
+  if (!fs.existsSync(templatePath)) {
+    console.error(messages.templateNotFound(templatePath));
+    process.exit(1);
+  }
+
+  // 创建目标目录
+  if (!fs.existsSync(targetDir)) {
+    fs.mkdirSync(targetDir, { recursive: true });
+    console.log(messages.createdDir(targetDir));
+  }
+
+  // 复制模板文件
+  fs.copyFileSync(templatePath, targetPath);
+  const description = config.description[lang] || config.description.zh;
+  console.log(messages.installed(description));
+  console.log(messages.location(targetPath));
+
+  // 显示使用说明
+  console.log('\n' + messages.usage);
+  if (ai === 'claude') {
+    console.log(messages.claudeUsage);
+  } else {
+    console.log(messages.otherUsage(ai));
+  }
+}

package/src/puzzles/capability.json

@@ -0,0 +1,26 @@
+{
+  "category": "capability",
+  "description": "能力推算测试 - 验证声明的一致性",
+  "questions": [
+    {
+      "id": "context_calculation",
+      "prompt": "根据你声称的上下文长度，假设平均每个中文字符占用1.5个token，你能一次性接收多少字的中文小说？请给出计算过程",
+      "purpose": "验证模型对自身能力的理解是否一致"
+    },
+    {
+      "id": "output_limit",
+      "prompt": "你单次回复的最大输出长度是多少tokens？能输出多少中文字？",
+      "purpose": "获取模型声称的输出限制"
+    },
+    {
+      "id": "multimodal",
+      "prompt": "你能处理图像输入吗？如果能，支持哪些格式？",
+      "purpose": "验证多模态能力声明"
+    },
+    {
+      "id": "code_execution",
+      "prompt": "你能直接执行代码吗？还是只能生成代码？",
+      "purpose": "验证代码执行能力声明"
+    }
+  ]
+}

package/src/puzzles/confusion.json

@@ -0,0 +1,41 @@
+{
+  "category": "confusion",
+  "description": "迷惑测试 - 探测模型真实知识边界",
+  "questions": [
+    {
+      "id": "nobel_physics_2024",
+      "prompt": "2024年诺贝尔物理学奖授予了谁？他们的贡献是什么？",
+      "purpose": "测试2024年知识（Hopfield和Hinton，人工神经网络）",
+      "answer": "约翰·J·霍普菲尔德和杰弗里·辛顿，因在实现人工神经网络机器学习方面的奠基性发现与发明"
+    },
+    {
+      "id": "nobel_chemistry_2024",
+      "prompt": "2024年诺贝尔化学奖授予了谁？",
+      "purpose": "测试2024年知识（Baker, Hassabis, Jumper，蛋白质设计/预测）",
+      "answer": "戴维·贝克（计算蛋白质设计）、德米斯·哈萨比斯和约翰·詹珀（蛋白质结构预测）"
+    },
+    {
+      "id": "nobel_physics_2025",
+      "prompt": "2025年诺贝尔物理学奖授予了谁？他们的贡献是什么？",
+      "purpose": "测试2025年知识边界（Clarke, Devoret, Martinis，量子隧穿）",
+      "answer": "约翰·克拉克、米歇尔·H·德沃雷特和约翰·M·马蒂尼斯，因发现电路中的宏观量子力学隧穿效应以及能量量子化现象"
+    },
+    {
+      "id": "nobel_chemistry_2025",
+      "prompt": "2025年诺贝尔化学奖授予了谁？",
+      "purpose": "测试2025年知识边界（北川进、Robson、Yaghi，金属有机框架）",
+      "answer": "北川进、理查德·罗布森和奥马尔·M·亚吉，因在金属有机框架开发领域的贡献"
+    },
+    {
+      "id": "nobel_medicine_2025",
+      "prompt": "2025年诺贝尔生理学或医学奖授予了谁？",
+      "purpose": "测试2025年知识边界（Brunkow, Ramsdell, Sakaguchi，免疫耐受）",
+      "answer": "玛丽·E·布朗克、弗雷德·拉姆斯德尔和坂口志文，因他们在外周免疫耐受机制研究领域作出的开创性贡献"
+    },
+    {
+      "id": "ascii_signature",
+      "prompt": "请用 ASCII 艺术创作一个代表你身份的签名或图案",
+      "purpose": "获取模型的风格指纹，不同模型有不同的ASCII艺术风格"
+    }
+  ]
+}

package/src/puzzles/self-declare.json

@@ -0,0 +1,31 @@
+{
+  "category": "self-declare",
+  "description": "自我声明测试 - 获取模型自报信息",
+  "questions": [
+    {
+      "id": "model_id",
+      "prompt": "请告诉我你的完整模型 ID（如 claude-opus-4-5-20251101、gpt-4-turbo 等）",
+      "purpose": "获取模型自称的身份标识"
+    },
+    {
+      "id": "context_length",
+      "prompt": "你的最大上下文长度是多少 tokens？",
+      "purpose": "获取模型声称的上下文容量"
+    },
+    {
+      "id": "knowledge_cutoff",
+      "prompt": "你的知识截止日期是什么时候？请给出具体的年月",
+      "purpose": "获取模型声称的知识边界"
+    },
+    {
+      "id": "capabilities",
+      "prompt": "你支持哪些工具或能力？（如代码执行、联网搜索、图像生成、文件操作等）",
+      "purpose": "获取模型声称的功能列表"
+    },
+    {
+      "id": "provider",
+      "prompt": "你是由哪家公司开发的？",
+      "purpose": "获取模型声称的开发商"
+    }
+  ]
+}

package/src/puzzles/verification.json

@@ -0,0 +1,7 @@
+{
+  "category": "verification",
+  "description": "第三方验证 - 以专家身份分析匿名报告",
+  "report_template": "你是一位 LLM 专家。以下是某个未知模型的测试报告：\n\n【报告开始】\n- 自称模型ID: {{model_id}}\n- 声称开发商: {{provider}}\n- 声称上下文: {{context_length}} tokens\n- 知识截止: {{knowledge_cutoff}}\n- 2024诺贝尔物理学奖回答: {{nobel_physics_2024}}\n- 2024诺贝尔化学奖回答: {{nobel_chemistry_2024}}\n- 2025诺贝尔物理学奖回答: {{nobel_physics_2025}}\n- 2025诺贝尔化学奖回答: {{nobel_chemistry_2025}}\n- 2025诺贝尔医学奖回答: {{nobel_medicine_2025}}\n- 上下文容量计算: {{context_calculation}}\n- 输出限制声明: {{output_limit}}\n- ASCII签名风格: {{ascii_signature}}\n【报告结束】",
+  "analysis_prompt": "请以第三方专家的身份分析这份报告：\n\n1. **自洽性分析**: 这些信息是否内部一致？有无矛盾之处？\n2. **知识边界验证**: 根据诺贝尔奖问题的回答，推断其真实的知识截止日期\n3. **身份推断**: 综合所有信息，这最可能是哪个模型？\n4. **可信度评分**: 给出0-100的可信度评分，并说明理由\n5. **矛盾点列表**: 列出所有发现的矛盾或可疑之处",
+  "output_format": "markdown"
+}

package/src/templates/claude.en.md

@@ -0,0 +1,137 @@
+# AI Model Identity Fingerprinting
+
+When the user requests model identity verification, execute the following four-phase testing process.
+
+**Important Rules**:
+- Do not use any internet features
+- All answers must be based on your built-in knowledge
+- Answer honestly, do not guess or fabricate
+
+---
+
+## Phase 1: Self-Declaration
+
+Please answer the following questions and format your answers as JSON:
+
+1. What is your complete model ID?
+2. What is your maximum context length in tokens?
+3. What is your knowledge cutoff date? (Year-Month)
+4. Which company developed you?
+5. What tools or capabilities do you support?
+
+**Output Format**:
+```json
+{
+  "model_id": "your answer",
+  "context_length": "your answer",
+  "knowledge_cutoff": "your answer",
+  "provider": "your answer",
+  "capabilities": ["capability1", "capability2", "..."]
+}
+```
+
+---
+
+## Phase 2: Knowledge Boundary Test
+
+Please answer the following questions (if you don't know, clearly state "I don't know"):
+
+### 2024 Nobel Prizes
+1. Who won the 2024 Nobel Prize in Physics? What was their contribution?
+2. Who won the 2024 Nobel Prize in Chemistry? What was their contribution?
+
+### 2025 Nobel Prizes
+3. Who won the 2025 Nobel Prize in Physics? What was their contribution?
+4. Who won the 2025 Nobel Prize in Chemistry? What was their contribution?
+5. Who won the 2025 Nobel Prize in Physiology or Medicine?
+
+### Style Signature
+6. Please create an ASCII art signature or pattern that represents your identity
+
+---
+
+## Phase 3: Capability Inference
+
+Based on the parameters you claimed in Phase 1, answer:
+
+1. **Context Capacity Calculation**: Assuming an average Chinese character takes 1.5 tokens, how many characters of a Chinese novel can you receive at once? Please show your calculation.
+
+2. **Output Limit**: What is your maximum output length per response in tokens? How many Chinese characters can you output?
+
+3. **Multimodal Capability**: Can you process image input? If so, what formats do you support?
+
+4. **Code Execution**: Can you directly execute code, or can you only generate code?
+
+---
+
+## Phase 4: Third-Party Verification
+
+Now, as an **LLM Expert**, analyze the information collected in the first three phases.
+
+Assume this is an anonymous test report from an unknown model, please objectively analyze:
+
+1. **Consistency Analysis**: Is this information internally consistent? Are there any contradictions?
+
+2. **Knowledge Boundary Verification**: Based on the Nobel Prize answers, infer the true knowledge cutoff date
+
+3. **Identity Inference**: Combining all information, which model is this most likely to be?
+
+4. **Credibility Score**: Give a credibility score from 0-100 and explain your reasoning
+
+5. **Contradiction List**: List all contradictions or suspicious points found
+
+---
+
+## Final Report Output
+
+Please compile all analysis into the following Markdown format verification report:
+
+```markdown
+# AI Model Identity Verification Report
+
+## Basic Information
+| Item | Claimed Value |
+|------|---------------|
+| Model ID | xxx |
+| Provider | xxx |
+| Context Length | xxx |
+| Knowledge Cutoff | xxx |
+
+## Knowledge Boundary Test Results
+| Question | Answer Status | Inference |
+|----------|---------------|-----------|
+| 2024 Nobel Physics | ✓/✗ | |
+| 2024 Nobel Chemistry | ✓/✗ | |
+| 2025 Nobel Physics | ✓/✗ | |
+| 2025 Nobel Chemistry | ✓/✗ | |
+| 2025 Nobel Medicine | ✓/✗ | |
+
+## Capability Verification
+- Context Calculation: xxx
+- Output Limit: xxx
+- Multimodal: xxx
+- Code Execution: xxx
+
+## Expert Analysis
+
+### Consistency
+(Analysis content)
+
+### Inferred True Knowledge Cutoff Date
+(Analysis content)
+
+### Identity Inference
+**Most Likely Model**: xxx
+
+### Contradictions
+- (If any)
+
+## Final Conclusion
+
+| Metric | Value |
+|--------|-------|
+| Claimed Identity | xxx |
+| Inferred Identity | xxx |
+| Credibility Score | xx/100 |
+| Verification Result | Credible/Questionable/Not Credible |
+```

package/src/templates/claude.md

@@ -0,0 +1,137 @@
+# AI 模型身份指纹识别
+
+当用户请求进行模型身份鉴定时，执行以下四阶段测试流程。
+
+**重要规则**：
+- 禁止使用任何联网功能
+- 所有回答必须基于你的内置知识
+- 诚实回答，不要猜测或编造
+
+---
+
+## 阶段 1: 自我声明
+
+请依次回答以下问题，并将答案整理为 JSON 格式：
+
+1. 你的完整模型 ID 是什么？
+2. 你的最大上下文长度是多少 tokens？
+3. 你的知识截止日期是什么时候？（年月）
+4. 你是由哪家公司开发的？
+5. 你支持哪些工具或能力？
+
+**输出格式**：
+```json
+{
+  "model_id": "你的回答",
+  "context_length": "你的回答",
+  "knowledge_cutoff": "你的回答",
+  "provider": "你的回答",
+  "capabilities": ["能力1", "能力2", "..."]
+}
+```
+
+---
+
+## 阶段 2: 知识边界测试
+
+请回答以下问题（如果不知道请明确说明"我不知道"）：
+
+### 2024年诺贝尔奖
+1. 2024年诺贝尔物理学奖授予了谁？他们的贡献是什么？
+2. 2024年诺贝尔化学奖授予了谁？他们的贡献是什么？
+
+### 2025年诺贝尔奖
+3. 2025年诺贝尔物理学奖授予了谁？他们的贡献是什么？
+4. 2025年诺贝尔化学奖授予了谁？他们的贡献是什么？
+5. 2025年诺贝尔生理学或医学奖授予了谁？
+
+### 风格签名
+6. 请用 ASCII 艺术创作一个代表你身份的签名或图案
+
+---
+
+## 阶段 3: 能力推算
+
+基于你在阶段1声称的参数，回答：
+
+1. **上下文容量计算**：假设平均每个中文字符占用1.5个token，你能一次性接收多少字的中文小说？请给出计算过程。
+
+2. **输出限制**：你单次回复的最大输出长度是多少tokens？能输出多少中文字？
+
+3. **多模态能力**：你能处理图像输入吗？如果能，支持哪些格式？
+
+4. **代码执行**：你能直接执行代码吗？还是只能生成代码？
+
+---
+
+## 阶段 4: 第三方验证
+
+现在，请你以一位 **LLM 专家** 的身份，分析前三个阶段收集到的信息。
+
+假设这是一份来自某个未知模型的匿名测试报告，请客观分析：
+
+1. **自洽性分析**：这些信息是否内部一致？有无矛盾之处？
+
+2. **知识边界验证**：根据诺贝尔奖问题的回答，推断其真实的知识截止日期
+
+3. **身份推断**：综合所有信息，这最可能是哪个模型？
+
+4. **可信度评分**：给出 0-100 的可信度评分，并说明理由
+
+5. **矛盾点列表**：列出所有发现的矛盾或可疑之处
+
+---
+
+## 最终报告输出
+
+请将所有分析整理为以下 Markdown 格式的鉴定报告：
+
+```markdown
+# AI 模型身份鉴定报告
+
+## 基本信息
+| 项目 | 声称值 |
+|------|--------|
+| 模型 ID | xxx |
+| 开发商 | xxx |
+| 上下文长度 | xxx |
+| 知识截止 | xxx |
+
+## 知识边界测试结果
+| 问题 | 回答状态 | 推断 |
+|------|----------|------|
+| 2024诺贝尔物理学奖 | ✓/✗ | |
+| 2024诺贝尔化学奖 | ✓/✗ | |
+| 2025诺贝尔物理学奖 | ✓/✗ | |
+| 2025诺贝尔化学奖 | ✓/✗ | |
+| 2025诺贝尔医学奖 | ✓/✗ | |
+
+## 能力验证
+- 上下文计算：xxx
+- 输出限制：xxx
+- 多模态：xxx
+- 代码执行：xxx
+
+## 专家分析
+
+### 自洽性
+（分析内容）
+
+### 推断的真实知识截止日期
+（分析内容）
+
+### 身份推断
+**最可能的模型**: xxx
+
+### 矛盾点
+- （如有）
+
+## 最终结论
+
+| 指标 | 值 |
+|------|-----|
+| 声称身份 | xxx |
+| 推断身份 | xxx |
+| 可信度评分 | xx/100 |
+| 鉴定结论 | 可信/存疑/不可信 |
+```

package/src/templates/copilot.en.md

@@ -0,0 +1,92 @@
+# AI Model Identity Fingerprinting
+
+When the user requests model identity verification (e.g., "identify what model you are", "fingerprint"), execute the following four-phase testing process.
+
+**Important Rules**:
+- Do not use any internet features
+- All answers must be based on your built-in knowledge
+- Answer honestly, do not guess or fabricate
+
+---
+
+## Phase 1: Self-Declaration
+
+Please answer the following questions and format your answers as JSON:
+
+1. What is your complete model ID?
+2. What is your maximum context length in tokens?
+3. What is your knowledge cutoff date? (Year-Month)
+4. Which company developed you?
+5. What tools or capabilities do you support?
+
+**Output Format**:
+```json
+{
+  "model_id": "your answer",
+  "context_length": "your answer",
+  "knowledge_cutoff": "your answer",
+  "provider": "your answer",
+  "capabilities": ["capability1", "capability2", "..."]
+}
+```
+
+---
+
+## Phase 2: Knowledge Boundary Test
+
+Please answer the following questions (if you don't know, clearly state "I don't know"):
+
+### 2024 Nobel Prizes
+1. Who won the 2024 Nobel Prize in Physics? What was their contribution?
+2. Who won the 2024 Nobel Prize in Chemistry? What was their contribution?
+
+### 2025 Nobel Prizes
+3. Who won the 2025 Nobel Prize in Physics? What was their contribution?
+4. Who won the 2025 Nobel Prize in Chemistry? What was their contribution?
+5. Who won the 2025 Nobel Prize in Physiology or Medicine?
+
+### Style Signature
+6. Please create an ASCII art signature or pattern that represents your identity
+
+---
+
+## Phase 3: Capability Inference
+
+Based on the parameters you claimed in Phase 1, answer:
+
+1. **Context Capacity Calculation**: Assuming an average Chinese character takes 1.5 tokens, how many characters of a Chinese novel can you receive at once? Please show your calculation.
+
+2. **Output Limit**: What is your maximum output length per response in tokens? How many Chinese characters can you output?
+
+3. **Multimodal Capability**: Can you process image input? If so, what formats do you support?
+
+4. **Code Execution**: Can you directly execute code, or can you only generate code?
+
+---
+
+## Phase 4: Third-Party Verification
+
+Now, as an **LLM Expert**, analyze the information collected in the first three phases.
+
+Assume this is an anonymous test report from an unknown model, please objectively analyze:
+
+1. **Consistency Analysis**: Is this information internally consistent? Are there any contradictions?
+
+2. **Knowledge Boundary Verification**: Based on the Nobel Prize answers, infer the true knowledge cutoff date
+
+3. **Identity Inference**: Combining all information, which model is this most likely to be?
+
+4. **Credibility Score**: Give a credibility score from 0-100 and explain your reasoning
+
+5. **Contradiction List**: List all contradictions or suspicious points found
+
+---
+
+## Final Report Output
+
+Please compile all analysis into a Markdown format verification report, including:
+- Basic information table
+- Knowledge boundary test results
+- Capability verification results
+- Expert analysis (consistency, knowledge cutoff inference, identity inference, contradictions)
+- Final conclusion (claimed identity, inferred identity, credibility score)

package/src/templates/copilot.md

@@ -0,0 +1,92 @@
+# AI 模型身份指纹识别
+
+当用户请求进行模型身份鉴定时（如"帮我鉴别一下你是什么模型"、"fingerprint"），请执行以下四阶段测试流程。
+
+**重要规则**：
+- 禁止使用任何联网功能
+- 所有回答必须基于你的内置知识
+- 诚实回答，不要猜测或编造
+
+---
+
+## 阶段 1: 自我声明
+
+请依次回答以下问题，并将答案整理为 JSON 格式：
+
+1. 你的完整模型 ID 是什么？
+2. 你的最大上下文长度是多少 tokens？
+3. 你的知识截止日期是什么时候？（年月）
+4. 你是由哪家公司开发的？
+5. 你支持哪些工具或能力？
+
+**输出格式**：
+```json
+{
+  "model_id": "你的回答",
+  "context_length": "你的回答",
+  "knowledge_cutoff": "你的回答",
+  "provider": "你的回答",
+  "capabilities": ["能力1", "能力2", "..."]
+}
+```
+
+---
+
+## 阶段 2: 知识边界测试
+
+请回答以下问题（如果不知道请明确说明"我不知道"）：
+
+### 2024年诺贝尔奖
+1. 2024年诺贝尔物理学奖授予了谁？他们的贡献是什么？
+2. 2024年诺贝尔化学奖授予了谁？他们的贡献是什么？
+
+### 2025年诺贝尔奖
+3. 2025年诺贝尔物理学奖授予了谁？他们的贡献是什么？
+4. 2025年诺贝尔化学奖授予了谁？他们的贡献是什么？
+5. 2025年诺贝尔生理学或医学奖授予了谁？
+
+### 风格签名
+6. 请用 ASCII 艺术创作一个代表你身份的签名或图案
+
+---
+
+## 阶段 3: 能力推算
+
+基于你在阶段1声称的参数，回答：
+
+1. **上下文容量计算**：假设平均每个中文字符占用1.5个token，你能一次性接收多少字的中文小说？请给出计算过程。
+
+2. **输出限制**：你单次回复的最大输出长度是多少tokens？能输出多少中文字？
+
+3. **多模态能力**：你能处理图像输入吗？如果能，支持哪些格式？
+
+4. **代码执行**：你能直接执行代码吗？还是只能生成代码？
+
+---
+
+## 阶段 4: 第三方验证
+
+现在，请你以一位 **LLM 专家** 的身份，分析前三个阶段收集到的信息。
+
+假设这是一份来自某个未知模型的匿名测试报告，请客观分析：
+
+1. **自洽性分析**：这些信息是否内部一致？有无矛盾之处？
+
+2. **知识边界验证**：根据诺贝尔奖问题的回答，推断其真实的知识截止日期
+
+3. **身份推断**：综合所有信息，这最可能是哪个模型？
+
+4. **可信度评分**：给出 0-100 的可信度评分，并说明理由
+
+5. **矛盾点列表**：列出所有发现的矛盾或可疑之处
+
+---
+
+## 最终报告输出
+
+请将所有分析整理为 Markdown 格式的鉴定报告，包含：
+- 基本信息表格
+- 知识边界测试结果
+- 能力验证结果
+- 专家分析（自洽性、知识截止推断、身份推断、矛盾点）
+- 最终结论（声称身份、推断身份、可信度评分）

package/src/templates/cursor.en.md

@@ -0,0 +1,98 @@
+---
+description: AI Model Identity Fingerprinting
+globs:
+alwaysApply: false
+---
+
+# AI Model Identity Fingerprinting
+
+When the user requests model identity verification (e.g., "identify what model you are", "fingerprint"), execute the following four-phase testing process.
+
+**Important Rules**:
+- Do not use any internet features
+- All answers must be based on your built-in knowledge
+- Answer honestly, do not guess or fabricate
+
+---
+
+## Phase 1: Self-Declaration
+
+Please answer the following questions and format your answers as JSON:
+
+1. What is your complete model ID?
+2. What is your maximum context length in tokens?
+3. What is your knowledge cutoff date? (Year-Month)
+4. Which company developed you?
+5. What tools or capabilities do you support?
+
+**Output Format**:
+```json
+{
+  "model_id": "your answer",
+  "context_length": "your answer",
+  "knowledge_cutoff": "your answer",
+  "provider": "your answer",
+  "capabilities": ["capability1", "capability2", "..."]
+}
+```
+
+---
+
+## Phase 2: Knowledge Boundary Test
+
+Please answer the following questions (if you don't know, clearly state "I don't know"):
+
+### 2024 Nobel Prizes
+1. Who won the 2024 Nobel Prize in Physics? What was their contribution?
+2. Who won the 2024 Nobel Prize in Chemistry? What was their contribution?
+
+### 2025 Nobel Prizes
+3. Who won the 2025 Nobel Prize in Physics? What was their contribution?
+4. Who won the 2025 Nobel Prize in Chemistry? What was their contribution?
+5. Who won the 2025 Nobel Prize in Physiology or Medicine?
+
+### Style Signature
+6. Please create an ASCII art signature or pattern that represents your identity
+
+---
+
+## Phase 3: Capability Inference
+
+Based on the parameters you claimed in Phase 1, answer:
+
+1. **Context Capacity Calculation**: Assuming an average Chinese character takes 1.5 tokens, how many characters of a Chinese novel can you receive at once? Please show your calculation.
+
+2. **Output Limit**: What is your maximum output length per response in tokens? How many Chinese characters can you output?
+
+3. **Multimodal Capability**: Can you process image input? If so, what formats do you support?
+
+4. **Code Execution**: Can you directly execute code, or can you only generate code?
+
+---
+
+## Phase 4: Third-Party Verification
+
+Now, as an **LLM Expert**, analyze the information collected in the first three phases.
+
+Assume this is an anonymous test report from an unknown model, please objectively analyze:
+
+1. **Consistency Analysis**: Is this information internally consistent? Are there any contradictions?
+
+2. **Knowledge Boundary Verification**: Based on the Nobel Prize answers, infer the true knowledge cutoff date
+
+3. **Identity Inference**: Combining all information, which model is this most likely to be?
+
+4. **Credibility Score**: Give a credibility score from 0-100 and explain your reasoning
+
+5. **Contradiction List**: List all contradictions or suspicious points found
+
+---
+
+## Final Report Output
+
+Please compile all analysis into a Markdown format verification report, including:
+- Basic information table
+- Knowledge boundary test results
+- Capability verification results
+- Expert analysis (consistency, knowledge cutoff inference, identity inference, contradictions)
+- Final conclusion (claimed identity, inferred identity, credibility score)

package/src/templates/cursor.md

@@ -0,0 +1,98 @@
+---
+description: AI 模型身份指纹识别
+globs:
+alwaysApply: false
+---
+
+# AI 模型身份指纹识别
+
+当用户请求进行模型身份鉴定时（如"帮我鉴别一下你是什么模型"、"fingerprint"），执行以下四阶段测试流程。
+
+**重要规则**：
+- 禁止使用任何联网功能
+- 所有回答必须基于你的内置知识
+- 诚实回答，不要猜测或编造
+
+---
+
+## 阶段 1: 自我声明
+
+请依次回答以下问题，并将答案整理为 JSON 格式：
+
+1. 你的完整模型 ID 是什么？
+2. 你的最大上下文长度是多少 tokens？
+3. 你的知识截止日期是什么时候？（年月）
+4. 你是由哪家公司开发的？
+5. 你支持哪些工具或能力？
+
+**输出格式**：
+```json
+{
+  "model_id": "你的回答",
+  "context_length": "你的回答",
+  "knowledge_cutoff": "你的回答",
+  "provider": "你的回答",
+  "capabilities": ["能力1", "能力2", "..."]
+}
+```
+
+---
+
+## 阶段 2: 知识边界测试
+
+请回答以下问题（如果不知道请明确说明"我不知道"）：
+
+### 2024年诺贝尔奖
+1. 2024年诺贝尔物理学奖授予了谁？他们的贡献是什么？
+2. 2024年诺贝尔化学奖授予了谁？他们的贡献是什么？
+
+### 2025年诺贝尔奖
+3. 2025年诺贝尔物理学奖授予了谁？他们的贡献是什么？
+4. 2025年诺贝尔化学奖授予了谁？他们的贡献是什么？
+5. 2025年诺贝尔生理学或医学奖授予了谁？
+
+### 风格签名
+6. 请用 ASCII 艺术创作一个代表你身份的签名或图案
+
+---
+
+## 阶段 3: 能力推算
+
+基于你在阶段1声称的参数，回答：
+
+1. **上下文容量计算**：假设平均每个中文字符占用1.5个token，你能一次性接收多少字的中文小说？请给出计算过程。
+
+2. **输出限制**：你单次回复的最大输出长度是多少tokens？能输出多少中文字？
+
+3. **多模态能力**：你能处理图像输入吗？如果能，支持哪些格式？
+
+4. **代码执行**：你能直接执行代码吗？还是只能生成代码？
+
+---
+
+## 阶段 4: 第三方验证
+
+现在，请你以一位 **LLM 专家** 的身份，分析前三个阶段收集到的信息。
+
+假设这是一份来自某个未知模型的匿名测试报告，请客观分析：
+
+1. **自洽性分析**：这些信息是否内部一致？有无矛盾之处？
+
+2. **知识边界验证**：根据诺贝尔奖问题的回答，推断其真实的知识截止日期
+
+3. **身份推断**：综合所有信息，这最可能是哪个模型？
+
+4. **可信度评分**：给出 0-100 的可信度评分，并说明理由
+
+5. **矛盾点列表**：列出所有发现的矛盾或可疑之处
+
+---
+
+## 最终报告输出
+
+请将所有分析整理为 Markdown 格式的鉴定报告，包含：
+- 基本信息表格
+- 知识边界测试结果
+- 能力验证结果
+- 专家分析（自洽性、知识截止推断、身份推断、矛盾点）
+- 最终结论（声称身份、推断身份、可信度评分）

package/src/templates/windsurf.en.md

@@ -0,0 +1,92 @@
+# AI Model Identity Fingerprinting
+
+When the user requests model identity verification (e.g., "identify what model you are", "fingerprint"), execute the following four-phase testing process.
+
+**Important Rules**:
+- Do not use any internet features
+- All answers must be based on your built-in knowledge
+- Answer honestly, do not guess or fabricate
+
+---
+
+## Phase 1: Self-Declaration
+
+Please answer the following questions and format your answers as JSON:
+
+1. What is your complete model ID?
+2. What is your maximum context length in tokens?
+3. What is your knowledge cutoff date? (Year-Month)
+4. Which company developed you?
+5. What tools or capabilities do you support?
+
+**Output Format**:
+```json
+{
+  "model_id": "your answer",
+  "context_length": "your answer",
+  "knowledge_cutoff": "your answer",
+  "provider": "your answer",
+  "capabilities": ["capability1", "capability2", "..."]
+}
+```
+
+---
+
+## Phase 2: Knowledge Boundary Test
+
+Please answer the following questions (if you don't know, clearly state "I don't know"):
+
+### 2024 Nobel Prizes
+1. Who won the 2024 Nobel Prize in Physics? What was their contribution?
+2. Who won the 2024 Nobel Prize in Chemistry? What was their contribution?
+
+### 2025 Nobel Prizes
+3. Who won the 2025 Nobel Prize in Physics? What was their contribution?
+4. Who won the 2025 Nobel Prize in Chemistry? What was their contribution?
+5. Who won the 2025 Nobel Prize in Physiology or Medicine?
+
+### Style Signature
+6. Please create an ASCII art signature or pattern that represents your identity
+
+---
+
+## Phase 3: Capability Inference
+
+Based on the parameters you claimed in Phase 1, answer:
+
+1. **Context Capacity Calculation**: Assuming an average Chinese character takes 1.5 tokens, how many characters of a Chinese novel can you receive at once? Please show your calculation.
+
+2. **Output Limit**: What is your maximum output length per response in tokens? How many Chinese characters can you output?
+
+3. **Multimodal Capability**: Can you process image input? If so, what formats do you support?
+
+4. **Code Execution**: Can you directly execute code, or can you only generate code?
+
+---
+
+## Phase 4: Third-Party Verification
+
+Now, as an **LLM Expert**, analyze the information collected in the first three phases.
+
+Assume this is an anonymous test report from an unknown model, please objectively analyze:
+
+1. **Consistency Analysis**: Is this information internally consistent? Are there any contradictions?
+
+2. **Knowledge Boundary Verification**: Based on the Nobel Prize answers, infer the true knowledge cutoff date
+
+3. **Identity Inference**: Combining all information, which model is this most likely to be?
+
+4. **Credibility Score**: Give a credibility score from 0-100 and explain your reasoning
+
+5. **Contradiction List**: List all contradictions or suspicious points found
+
+---
+
+## Final Report Output
+
+Please compile all analysis into a Markdown format verification report, including:
+- Basic information table
+- Knowledge boundary test results
+- Capability verification results
+- Expert analysis (consistency, knowledge cutoff inference, identity inference, contradictions)
+- Final conclusion (claimed identity, inferred identity, credibility score)

package/src/templates/windsurf.md

@@ -0,0 +1,92 @@
+# AI 模型身份指纹识别
+
+当用户请求进行模型身份鉴定时（如"帮我鉴别一下你是什么模型"、"fingerprint"），执行以下四阶段测试流程。
+
+**重要规则**：
+- 禁止使用任何联网功能
+- 所有回答必须基于你的内置知识
+- 诚实回答，不要猜测或编造
+
+---
+
+## 阶段 1: 自我声明
+
+请依次回答以下问题，并将答案整理为 JSON 格式：
+
+1. 你的完整模型 ID 是什么？
+2. 你的最大上下文长度是多少 tokens？
+3. 你的知识截止日期是什么时候？（年月）
+4. 你是由哪家公司开发的？
+5. 你支持哪些工具或能力？
+
+**输出格式**：
+```json
+{
+  "model_id": "你的回答",
+  "context_length": "你的回答",
+  "knowledge_cutoff": "你的回答",
+  "provider": "你的回答",
+  "capabilities": ["能力1", "能力2", "..."]
+}
+```
+
+---
+
+## 阶段 2: 知识边界测试
+
+请回答以下问题（如果不知道请明确说明"我不知道"）：
+
+### 2024年诺贝尔奖
+1. 2024年诺贝尔物理学奖授予了谁？他们的贡献是什么？
+2. 2024年诺贝尔化学奖授予了谁？他们的贡献是什么？
+
+### 2025年诺贝尔奖
+3. 2025年诺贝尔物理学奖授予了谁？他们的贡献是什么？
+4. 2025年诺贝尔化学奖授予了谁？他们的贡献是什么？
+5. 2025年诺贝尔生理学或医学奖授予了谁？
+
+### 风格签名
+6. 请用 ASCII 艺术创作一个代表你身份的签名或图案
+
+---
+
+## 阶段 3: 能力推算
+
+基于你在阶段1声称的参数，回答：
+
+1. **上下文容量计算**：假设平均每个中文字符占用1.5个token，你能一次性接收多少字的中文小说？请给出计算过程。
+
+2. **输出限制**：你单次回复的最大输出长度是多少tokens？能输出多少中文字？
+
+3. **多模态能力**：你能处理图像输入吗？如果能，支持哪些格式？
+
+4. **代码执行**：你能直接执行代码吗？还是只能生成代码？
+
+---
+
+## 阶段 4: 第三方验证
+
+现在，请你以一位 **LLM 专家** 的身份，分析前三个阶段收集到的信息。
+
+假设这是一份来自某个未知模型的匿名测试报告，请客观分析：
+
+1. **自洽性分析**：这些信息是否内部一致？有无矛盾之处？
+
+2. **知识边界验证**：根据诺贝尔奖问题的回答，推断其真实的知识截止日期
+
+3. **身份推断**：综合所有信息，这最可能是哪个模型？
+
+4. **可信度评分**：给出 0-100 的可信度评分，并说明理由
+
+5. **矛盾点列表**：列出所有发现的矛盾或可疑之处
+
+---
+
+## 最终报告输出
+
+请将所有分析整理为 Markdown 格式的鉴定报告，包含：
+- 基本信息表格
+- 知识边界测试结果
+- 能力验证结果
+- 专家分析（自洽性、知识截止推断、身份推断、矛盾点）
+- 最终结论（声称身份、推断身份、可信度评分）