← 返回 Skills 市场
166
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install vision-simulator
功能描述
Marketing strategy prediction with real-world data integration. Predicts campaign outcomes, product launches, and market reactions using multi-agent simulati...
安全使用建议
This skill is internally consistent for a marketing-simulation tool. Before installing, consider: (1) web_fetch will access public web data (news, social media, competitor sites) — confirm you are comfortable with the agent fetching and summarizing such content and check for any privacy or compliance constraints; (2) the SKILL.md mentions Pro real-time collection but does not declare any API keys — ask the publisher which credentials (if any) are required for Pro features and where those would be stored; (3) if you expect the agent to handle proprietary or sensitive inputs, verify how it treats and retains fetched data. If the SKILL.md later adds steps to read local files, environment variables, or download and execute code, that would materially raise risk and should be re-evaluated.
功能分析
Type: OpenClaw Skill
Name: vision-simulator
Version: 1.0.0
The 'vision-simulator' skill is a marketing strategy prediction tool that uses multi-agent simulation and real-world data integration. It utilizes standard OpenClaw capabilities such as `sessions_spawn` for agent orchestration and `web_fetch` for gathering market trends, both of which are directly aligned with its stated purpose in SKILL.md. There is no evidence of data exfiltration, malicious command execution, or harmful prompt injection.
能力评估
Purpose & Capability
Name/description (marketing prediction using live data + multi-agent simulation) match the SKILL.md: it describes keyword extraction, web data collection, and five agents. No unrelated binaries, env vars, or config paths are requested.
Instruction Scope
Instructions explicitly call for web_fetch (news, social media, competitor sites) and spawning five agent sessions and synthesizing results — this is within the stated purpose. Note: web_fetch could retrieve broad public web content (including public PII or sensitive competitor info) and the SKILL.md does not constrain which endpoints or how to handle sensitive data.
Install Mechanism
Instruction-only skill with no install spec and no code files — lowest-risk installation surface. No remote downloads or package installs are present.
Credentials
The skill requests no environment variables or credentials. It references data sources (Google News, Naver, social media) that might, in practice, require API keys for reliable access, but these are not declared — absence of declared credentials is not incoherent for a free/basic mode using public web fetches.
Persistence & Privilege
always is false, model invocation is enabled (normal). The skill doesn't request persistent system-wide privileges or modify other skills' configs.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install vision-simulator - 安装完成后，直接呼叫该 Skill 的名称或使用
/vision-simulator触发 - 根据 Skill 的参数说明提供必要输入，即可获得结构化输出
版本历史
v1.0.0
Initial release - Marketing prediction with real-world data integration, 5-agent simulation, success probability
元数据
常见问题
Vision Simulator 是什么？
Marketing strategy prediction with real-world data integration. Predicts campaign outcomes, product launches, and market reactions using multi-agent simulati... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 166 次。
如何安装 Vision Simulator？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install vision-simulator」即可一键安装，无需额外配置。
Vision Simulator 是免费的吗？
是的，Vision Simulator 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Vision Simulator 支持哪些平台？
Vision Simulator 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（cross-platform）。
谁开发了 Vision Simulator？
由 dorongss（@dorongss）开发并维护，当前版本 v1.0.0。
推荐 Skills