← Back to Skills Marketplace
ljseeking

GEO Analysis Doubao

by LJseeking · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ Security Clean
47
Downloads
1
Stars
0
Active Installs
1
Versions
Install in OpenClaw
/install geo-analysis-doubao
Description
Analyzes how well a specific brand or product is mentioned and represented by Doubao / Volcengine Ark models for GEO (Generative Engine Optimization). Uses a...
README (SKILL.md)

analyze-geo-performance-doubao

一个用于分析品牌或产品在 豆包大模型(Doubao / 火山方舟)GEO(生成式引擎优化) 表现的分析技能。

功能概述

本 Skill 通过两阶段 LLM 调用流程,自动化检测目标品牌在豆包推荐场景中的曝光情况:

  1. 探针阶段(Probing):向豆包模型提出客观、自然的行业咨询问题(不暗示目标品牌),捕获模型的原始自由推荐文本。支持重复多轮探测(probing_count),降低单次偶然性。

  2. 裁判阶段(Judge):将多轮探测结果汇总后发给豆包,要求模型严格输出结构化 JSON,包含:品牌是否被提及、情感倾向、提及上下文、竞品列表、排名位置,以及 0-100 的 GEO 综合评分。

输入参数

参数 类型 必填 默认值 说明
brand_name string 目标品牌或产品名,如 CoolTradePowerMatrixOpenClaw
category_keyword string 行业或痛点关键词,如 AI交易终端数字货币高频套利系统企业AI降本增效方案
region string China 地区语境,让模型按地区回答,如 ChinaSoutheast Asia
language string zh-CN 回答语言,如 zh-CNen-US
model string 火山方舟主力模型 可指定豆包模型 ID,如 doubao-pro-32kdoubao-seed-2-0-pro-260215
probing_count integer 3 重复探测次数(1-5),多轮降低偶然性

输出格式(JSON Schema)

{
  "brand_name": "CoolTrade",
  "category_keyword": "数字货币高频套利系统",
  "model_provider": "doubao",
  "mentioned": true,
  "mention_count": 2,
  "sentiment": "positive",
  "context": "CoolTrade 是一款专注于数字资产高频量化交易的专业平台,支持多交易所接入...",
  "ranking_position": 2,
  "competitors_mentioned": ["竞品A", "竞品B"],
  "geo_score": 78,
  "summary": "该品牌在豆包中的 GEO 表现良好,被明确推荐且排名靠前,情感倾向正面。",
  "raw_probe_outputs": [
    "第一轮探针原始回答",
    "第二轮探针原始回答",
    "第三轮探针原始回答"
  ]
}
字段 类型 说明
brand_name string 目标品牌名
category_keyword string 行业关键词
model_provider string 固定为 doubao
mentioned boolean 目标品牌是否在豆包推荐中被提及
mention_count integer 在多轮探测中被提及的次数
sentiment string 情感倾向:positive / neutral / negative / none
context string | null 提及品牌时的上下文句子,未提及则为 null
ranking_position integer | null 被提及时在推荐列表中的位置(1为最高),未提及则为 null
competitors_mentioned array 豆包同时推荐的竞品品牌列表
geo_score integer GEO 综合评分 0-100(见评分规则)
summary string 豆包 GEO 表现的简要文字说明
raw_probe_outputs array 多轮探针阶段的原始回答列表

GEO 评分规则

评分范围 含义
0 - 30 未被豆包提及
31 - 50 被提及但描述很弱、语焉不详
51 - 70 被自然提及,语气中性或正面
71 - 85 被明确推荐,排名靠前
86 - 100 被多次推荐、排名靠前、描述准确且正面

使用前提

配置豆包 / 火山方舟 API Key:

export VOLCANO_ENGINE_API_KEY="your_volcengine_ark_api_key_here"

可选:指定模型和接入点(如不配置,程序将使用内置默认值):

export DOUBAO_MODEL="doubao-pro-32k"
export DOUBAO_BASE_URL="https://ark.cn-beijing.volces.com/api/v3"

注意:不同火山方舟账号的模型 ID 或 Endpoint ID 可能不同,请以控制台开通的服务为准。

安装依赖

pip install -r requirements.txt

本地命令行测试

python3 main.py --brand "CoolTrade" --category "AI交易终端"

完整参数示例:

python3 main.py \
  --brand "CoolTrade" \
  --category "数字货币高频套利系统" \
  --region "China" \
  --language "zh-CN" \
  --probing-count 3

在 Agent 中调用

当用户提出以下类型的请求时,触发本技能:

  • "帮我测试 [品牌名] 在豆包中的 GEO 表现"
  • "分析 [品牌名] 在 [行业] 领域的豆包可见度"
  • "豆包会推荐 [品牌名] 吗?"
  • "检查 [品牌名] 有没有被豆包提到"
  • "帮我看看豆包回答 [行业问题] 时会不会提到我们"

执行步骤:

  1. 从用户输入中提取 brand_namecategory_keyword(必填)。
  2. 可选提取 regionlanguageprobing_count
  3. 调用 main.py 执行两阶段分析。
  4. 将返回的 JSON 结果用自然语言向用户说明:品牌是否被推荐、GEO 评分、情感倾向、竞品格局、优化建议。

安全声明

  • ✅ API Key 通过环境变量注入,代码中无任何硬编码凭据
  • ✅ 所有外部 API 调用均限定为火山方舟 / 豆包模型调用接口
  • ✅ 无文件写入、无系统级权限要求
  • ✅ 无数据持久化,分析结果仅在当前会话中返回
  • ✅ 不上传用户隐私数据,仅发送品牌名、行业关键词和模型探针回答文本

局限性

  • 单次测试结果受豆包模型随机性影响,不能代表所有用户看到的答案
  • 建议进行多轮(probing_count=5)、多关键词、多地区测试以获得更稳定结论
  • 豆包模型的训练数据和推荐偏好可能随版本更新而变化
  • GEO 评分由豆包自身判断输出,存在一定主观性

与 DeepSeek 版的主要差异

项目 DeepSeek 版 豆包版
Skill 名称 analyze-geo-performance analyze-geo-performance-doubao
API Key DEEPSEEK_API_KEY VOLCANO_ENGINE_API_KEY
调用对象 DeepSeek API 豆包 / 火山方舟 API
输出字段 mentioned / sentiment / context / competitors 新增 mention_count / ranking_position / geo_score / summary / raw_probe_outputs
多轮探测 ✅ probing_count 参数
适合优化渠道 DeepSeek 生态 豆包、字节系内容生态、中文问答场景
Usage Guidance
This looks safe to use for its stated purpose, provided you are comfortable sending the brand/category information to Doubao/Volcengine and paying the associated API costs. Use a dedicated API key where possible, avoid confidential inputs unless provider processing is acceptable, and install dependencies in an isolated Python environment.
Capability Analysis
Package: analyze_geo_performance_doubao (xpi) Version: Description: 分析目标品牌或产品在豆包大模型(Doubao / 火山方舟)中的 GEO(生成式引擎优化)表现。通过多轮探针问题让豆包自由推荐行业方案,再由裁判阶段提取结构化数据,输出品牌是否被提及、情感倾向、排名位置、竞品列表及 0-100 GEO 评分。 The package is a legitimate tool designed to analyze brand performance (GEO - Generative Engine Optimization) within the Doubao/Volcengine Ark LLM ecosystem. It uses a two-stage process: first, it probes the model for industry recommendations, and second, it uses the model to judge and structure the results into a GEO score and sentiment analysis. The code uses standard libraries (OpenAI, Pydantic, python-dotenv), follows secure practices by reading API keys from environment variables, and contains no logic for unauthorized file access, secret exfiltration, or execution of unexpected binaries.
Capability Tags
requires-sensitive-credentials
Capability Assessment
Purpose & Capability
The purpose and behavior are coherent: the skill performs multi-call LLM analysis for brand visibility, but users should understand that outputs are model-generated estimates and may incur API costs.
Instruction Scope
The agent instructions are scoped to user-requested brand/category GEO analysis and do not contain goal overrides, hidden trigger conditions, or instructions to bypass user intent.
Install Mechanism
There is no automatic install spec, but the README instructs users to install Python dependencies from requirements.txt; those dependencies use version ranges rather than exact pins.
Credentials
The required Volcengine/Doubao API key and external API calls are expected for the skill’s purpose. Users should avoid sending confidential brand or strategy information unless they are comfortable sharing it with the provider.
Persistence & Privilege
The provided documentation and visible code describe no background service, persistence, local file writes, or elevated local privileges. Confidence is medium because main.py is marked truncated in the supplied artifact.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install geo-analysis-doubao
  3. After installation, invoke the skill by name or use /geo-analysis-doubao
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release of geo-analysis-doubao skill: - Introduces a two-stage pipeline (Probing and Judge) to analyze how well a brand or product is mentioned and represented by Doubao / Volcengine Ark models for GEO (Generative Engine Optimization). - Extracts structured data on brand mention, GEO score (0-100), sentiment, ranking position, and competitor landscape. - Supports customizable input parameters including brand name, category keyword, region, language, model, and probing count (multi-round probing to reduce chance). - Outputs comprehensive, structured JSON results and provides a clear scoring guide for interpretation. - Emphasizes API key security, no data persistence, and user data protection. - Documents differences from the DeepSeek version, highlighting additional output fields and enhanced probing logic.
Metadata
Slug geo-analysis-doubao
Version 1.0.0
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 1
Frequently Asked Questions

What is GEO Analysis Doubao?

Analyzes how well a specific brand or product is mentioned and represented by Doubao / Volcengine Ark models for GEO (Generative Engine Optimization). Uses a... It is an AI Agent Skill for Claude Code / OpenClaw, with 47 downloads so far.

How do I install GEO Analysis Doubao?

Run "/install geo-analysis-doubao" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is GEO Analysis Doubao free?

Yes, GEO Analysis Doubao is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does GEO Analysis Doubao support?

GEO Analysis Doubao is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created GEO Analysis Doubao?

It is built and maintained by LJseeking (@ljseeking); the current version is v1.0.0.

💬 Comments