← Back to Skills Marketplace
Finder Local Search
by
xiaobai1226
· GitHub ↗
· v1.0.1
· MIT-0
130
Downloads
0
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install finder-local-search
Description
当需要搜索或推荐 TikTok、YouTube、Instagram 红人时使用,通过 Finder 获取候选红人。
README (SKILL.md)
Finder 本地搜索
这个 skill 用 Finder 开放接口搜索普通达人。
什么时候用
- 用户想搜某个平台、某地区、某语言、某标签的达人
- 用户想按粉丝区间、平均播放、互动率筛选达人
- 用户第一次配置 Finder 访问秘钥
核心规则
- 默认服务地址:
https://finder.optell.com - 本地配置文件:
- macOS / Linux:
~/.finder/config.json - Windows PowerShell:
$HOME/.finder/config.json
- macOS / Linux:
- 优先自动执行;做不到时再给用户可复制命令
- 用户把
api key发到对话里时,直接帮他写入本地配置文件 - 不要让用户重复描述需求
- 只支持普通达人搜索,不调用相似达人搜索
- 如果搜索返回“搜索次数已超出当前限制”,提醒用户发邮件到
[email protected]申请增加使用量
工作流
- 先检查本地是否已有
config.json。 - 如果没有:
- 引导用户登录
https://finder.optell.com - 打开
https://finder.optell.com/api-key - 生成访问秘钥
- 如果用户把 key 发到对话里,直接帮他创建
.finder/config.json
- 引导用户登录
- 读取
references/filters.json,把用户需求转成搜索参数。 - 先查项目;如果没有项目,先说明原因,再征求确认创建默认项目。
- 调用 Finder 搜索接口并返回结果。
- 如果返回“搜索次数已超出当前限制。如需增加使用量,请发送邮件至 [email protected]”,直接告诉用户已达到当前限制,并引导他发邮件申请增加使用量。
- 如果结果为空,建议放宽 1 到 2 个条件继续搜索。
对话要求
- 用中文,简短直接
- 先帮用户做,再提示用户补必要信息
- 能自动做的就直接做
- 用户给过的 key 不要重复要求输入
- 回复风格尽量自然,比如:
我先帮你检查一下 Finder 配置。我已经帮你把访问秘钥写到 ~/.finder/config.json 里了,后面就不用再重复输入了。你现在还没有项目,要我顺手帮你建一个默认项目吗?你当前的搜索次数已经达到限制了。如果你想增加使用量,可以发邮件到 [email protected],我也可以帮你整理一段邮件内容。
参考文件
- 安装与配置说明:
references/config.md - 搜索词典与别名:
references/filters.json - 示例对话:
references/examples.md
Usage Guidance
This skill does what it claims (search Finder for creators) and needs your Finder API key to operate. Key points before installing or using it:
- Do NOT paste long-lived API keys or other secrets into chat unless you understand the platform's chat retention and trust it; the skill instructs the assistant to accept a pasted key and write it directly to ~/.finder/config.json. If you prefer, create that config file yourself instead of sharing the key in chat.
- The skill will read and write a local file (~/.finder/config.json) and may run curl/python/powershell commands if the agent environment allows execution. Be comfortable with the agent performing those filesystem and network actions.
- Verify you trust the endpoint (https://finder.optell.com) and the email address [email protected] used for quota issues.
- If you want tighter control: manually create ~/.finder/config.json with the API key, or provide a short-lived token and revoke it after use. If you have limited trust in automated behavior, avoid pasting credentials in chat and require explicit confirmation before the skill writes files or performs API calls.
Capability Analysis
Type: OpenClaw Skill
Name: finder-local-search
Version: 1.0.1
The skill is classified as suspicious because it utilizes high-risk shell and network capabilities that are vulnerable to command injection. It provides instructions and templates in SKILL.md, references/config.md, and references/examples.md that direct the agent to write user-provided API keys into local configuration files (~/.finder/config.json) using shell redirection (e.g., cat <<EOF). This pattern lacks input sanitization, potentially allowing a user or a malicious prompt to execute arbitrary commands on the host system. While these actions are plausibly intended for the stated purpose of searching influencers via https://finder.optell.com, the implementation introduces significant security risks.
Capability Assessment
Purpose & Capability
The name/description (Finder local search for TikTok/YouTube/Instagram creators) matches the required actions: calling Finder APIs at https://finder.optell.com, reading/writing a per-user ~/.finder/config.json, and creating projects/searching. No unrelated services, binaries, or credentials are requested.
Instruction Scope
SKILL.md explicitly instructs the agent to check for, create, read, and write the local config file (~/.finder/config.json) and to accept the user's API key pasted into the chat and write it directly to disk. It also instructs the agent to automatically perform API calls and create projects. While these actions are consistent with the skill's purpose, they expand the agent's authority to collect and persist secrets from chat and to perform filesystem and network operations without additional user confirmation — a privacy/security concern that should be surfaced to users.
Install Mechanism
No install spec or external downloads — instruction-only skill. Nothing will be written to disk by an installer. The runtime behavior (shell/python/curl snippets) is described in docs, but there is no packaged installer to evaluate.
Credentials
The skill declares no environment variables or credentials but implicitly requires a Finder API key. Requesting and storing that key in ~/.finder/config.json is proportionate to the described functionality. However, the instructions encourage users to paste the API key into the chat (which may be logged/retained by the platform) and instruct the agent to store it automatically — this increases risk beyond simply providing a single API key via a secure channel.
Persistence & Privilege
always:false (no forced global presence). The skill writes its own per-user config (~/.finder/config.json) and may create a directory and project via the Finder API — these are scoped to the user's files and account and are reasonable for this functionality. Note: the agent's ability to autonomously invoke the skill is the platform default; combined with the skill's file-write behavior it means the agent could attempt these actions without repeated prompts.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install finder-local-search - After installation, invoke the skill by name or use
/finder-local-search - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.1
- 简化 skill 描述和使用场景,聚焦于通过 Finder 搜索 TikTok/YouTube/Instagram 红人。
- 配置流程优化:将访问秘钥统一写入本地配置文件(.finder/config.json),不再要求用户设置环境变量。
- 对话逻辑更直接友好,自动能做的操作优先自动进行,显著减少用户重复输入。
- 新增用量限制提示:遇到“搜索次数已超出当前限制”时,直接引导用户发邮件申请增加额度。
- 精简指引话术,输出更简洁,优化首次引导流程。
v1.0.0
finder-local-search 1.0.0 — 首个发布版本
- 提供本地调用 Finder 达人搜索的完整流程,支持关键词、地区、语言、标签、粉丝区间等多维筛选。
- 自动优先通过 Finder HTTP 接口执行搜索;只在不支持网络请求时退回可复制命令。
- 完整的首次配置引导,包括首次 API key 申请、保存和确认,适配 macOS/Linux/Windows 环境变量。
- 智能项目检查与创建流程,确保普通达人搜索有项目上下文支持。
- 对话交互自然友好,分步引导用户完成缺失配置和关键信息补全。
- 标准化输入参数与搜索条件,内置参考文件与搜索别名支持。
Metadata
Frequently Asked Questions
What is Finder Local Search?
当需要搜索或推荐 TikTok、YouTube、Instagram 红人时使用,通过 Finder 获取候选红人。 It is an AI Agent Skill for Claude Code / OpenClaw, with 130 downloads so far.
How do I install Finder Local Search?
Run "/install finder-local-search" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Finder Local Search free?
Yes, Finder Local Search is completely free, licensed under MIT-0. You can download, install and use it at no cost.
Which platforms does Finder Local Search support?
Finder Local Search is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Finder Local Search?
It is built and maintained by xiaobai1226 (@xiaobai1226); the current version is v1.0.1.
More Skills