Token导航 LogoToken导航

Agent Skill · 研究检索

thinkforce(thinkforce 搜索)

thinkforce 用于查找、检索和筛选相关信息,适合在 OpenClaw 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

thinkforce 是研究检索类 Agent Skill,可用宿主为 OpenClaw,通过ClawHub获取,权限为接触敏感数据;许可证为 MIT-0,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 22,321。

总安装 4,586周安装 195GitHub Stars 公开资料未说明下载 1,607核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
MIT-0
版本
1.1.0

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:thinkforce(thinkforce 搜索)
来源仓库:https://github.com/ade5791/thinkforce
安装命令:
openclaw skills install thinkforce
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 OpenClaw 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

ClawHubOpenClaw
openclaw skills install thinkforce

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

通过 REST API 将任务分派给您的 ThinkForce AI 代理,并轻松轮询结果,无需服务器设置或复杂的配置。

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 OpenClaw;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 openclaw skills install thinkforce;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要接触敏感数据,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

thinkforce 怎么安装?

在终端执行 openclaw skills install thinkforce。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

thinkforce 支持哪些 AI 工具?

公开资料标注支持 OpenClaw。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

thinkforce 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

thinkforce 是免费的吗?

来源标注的许可证是 MIT-0。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

thinkforce 用的人多吗?

公开数据是累计安装 4,586 次,近一周安装 195 次,下载 1,607 次。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

thinkforce 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 专注 OpenClaw 一种宿主,触发方式和使用习惯不用来回切换。
  • 许可证明确为 MIT-0,商用和分发边界好判断。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 OpenClaw,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。
  • 需要许可证清楚、并且有第三方扫描结果的团队项目。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
thinkforce

thinkforce 搜索 · 本页

4,586—接触敏感数据MIT-0通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

ThinkForce Skill

Dispatch tasks to your ThinkForce AI agent team at app.thinkforce.ai and poll for results — all via the REST API. No server setup needed.


Getting Your API Key

  1. Go to app.thinkforce.ai and sign up / log in.
  2. Open Settings → ThinkForce API.
  3. Click + Generate API Key — copy and save it.

Use your key as the X-TF-API-Key header on every request.


Quick Start

Step 1 — Get Your Company ID

Your API key automatically identifies your company:

GET https://app.thinkforce.ai/api/companies
X-TF-API-Key: tf_your_key_here

Response:

{
  "companyId": "your-company-id",
  "name": "Acme Corp",
  "status": "active",
  "agentCount": 4
}

Step 2 — List Your Agents

POST https://app.thinkforce.ai/api/agents
X-TF-API-Key: tf_your_key_here
Content-Type: application/json

{ "action": "list", "companyId": "your-company-id" }

Response:

[
  { "id": "abc123", "agentName": "Acme CEO", "agentRole": "CEO" },
  { "id": "def456", "agentName": "Dev", "agentRole": "Developer" }
]

Step 3 — Dispatch a Task

POST https://app.thinkforce.ai/api/agent-task
X-TF-API-Key: tf_your_key_here
Content-Type: application/json

{
  "companyID": "your-company-id",
  "targetAgentId": "abc123",
  "task": "Research the top 5 competitor apps and summarize their pricing."
}

Response:

{
  "taskId": "task-1773947855985-abc12",
  "status": "running"
}

Step 4 — Poll for Result

GET https://app.thinkforce.ai/api/agent-task?taskId=<taskId>&companyId=<companyId>
X-TF-API-Key: tf_your_key_here

Poll every 5–8 seconds. Most tasks complete in 15–90 seconds.

Completed response:

{
  "taskId": "task-1773947855985-abc12",
  "status": "complete",
  "result": "Here are the top 5 competitors..."
}

Full JavaScript Example

const API_KEY = 'tf_your_key_here';
const BASE = 'https://app.thinkforce.ai/api';

// 1. Get company
const co = await fetch(`${BASE}/companies`, {
  headers: { 'X-TF-API-Key': API_KEY }
}).then(r => r.json());

const companyID = co.companyId;

// 2. List agents to find the CEO
const agents = await fetch(`${BASE}/agents`, {
  method: 'POST',
  headers: { 'Content-Type': 'application/json', 'X-TF-API-Key': API_KEY },
  body: JSON.stringify({ action: 'list', companyId: companyID }),
}).then(r => r.json());

const ceo = agents.find(a => a.agentRole === 'CEO');

// 3. Dispatch task
const { taskId } = await fetch(`${BASE}/agent-task`, {
  method: 'POST',
  headers: { 'Content-Type': 'application/json', 'X-TF-API-Key': API_KEY },
  body: JSON.stringify({
    companyID,
    targetAgentId: ceo.id,
    task: 'Write a 30-day growth plan.',
  }),
}).then(r => r.json());

// 4. Poll for result
let result;
while (!result) {
  await new Promise(r => setTimeout(r, 6000));
  const data = await fetch(
    `${BASE}/agent-task?taskId=${taskId}&companyId=${companyID}`,
    { headers: { 'X-TF-API-Key': API_KEY } }
  ).then(r => r.json());

  if (data.result) result = data.result;
  if (data.status === 'error') throw new Error(data.error);
}

console.log(result);

Missions API

Missions are multi-step orchestrated workflows — create a mission, break it into subtasks, assign agents, and auto-execute the whole thing.

Create a Mission

POST https://app.thinkforce.ai/api/missions
X-TF-API-Key: tf_your_key_here
Content-Type: application/json

{
  "companyId": "your-company-id",
  "title": "Launch Q2 Growth Campaign",
  "description": "Research competitors, draft content plan, and build outreach strategy",
  "priority": "high",
  "createdBy": "user-id-optional"
}

Response:

{
  "id": "mission-abc123",
  "status": "planning",
  "title": "Launch Q2 Growth Campaign",
  ...
}

Get a Mission (with subtasks)

GET https://app.thinkforce.ai/api/missions/{missionId}?companyId=your-company-id
X-TF-API-Key: tf_your_key_here

Returns { mission, subtasks }.

List All Missions

GET https://app.thinkforce.ai/api/missions?companyId=your-company-id
X-TF-API-Key: tf_your_key_here

Update a Mission

PATCH https://app.thinkforce.ai/api/missions/{missionId}
X-TF-API-Key: tf_your_key_here
Content-Type: application/json

{
  "companyId": "your-company-id",
  "status": "active"
}

Status values: planning → active → completed | cancelled. Completing (status: "completed") triggers a mission completion email + Slack notification automatically.

Auto-Decompose into Subtasks

Let AI break the mission into 5–9 actionable subtasks automatically:

POST https://app.thinkforce.ai/api/missions/{missionId}/decompose
X-TF-API-Key: tf_your_key_here
Content-Type: application/json

{
  "title": "Launch Q2 Growth Campaign",
  "description": "Research competitors, draft content plan, and build outreach strategy"
}

Response:

{
  "subtasks": [
    { "title": "Research top 5 competitors", "workstationKey": "researching" },
    { "title": "Draft content calendar", "workstationKey": "working" },
    ...
  ]
}

Add a Subtask

POST https://app.thinkforce.ai/api/missions/{missionId}/subtasks
X-TF-API-Key: tf_your_key_here
Content-Type: application/json

{
  "companyId": "your-company-id",
  "title": "Research target audience",
  "workstationKey": "researching",
  "assignedAgentId": "agent-id-optional",
  "runInstructions": "Focus on 18-35 demographic in US markets"
}

workstationKey options: working | researching | syncing | error

Update / Delete a Subtask

PATCH https://app.thinkforce.ai/api/missions/{missionId}/subtasks/{subtaskId}
Content-Type: application/json

{
  "companyId": "your-company-id",
  "status": "done",
  "output": "Research complete. Key finding: ..."
}

For delete: DELETE same URL with companyId in body/query, or POST with { "action": "delete" }.

Run a Subtask (execute with assigned agent)

POST https://app.thinkforce.ai/api/missions/{missionId}/subtasks/{subtaskId}/run
X-TF-API-Key: tf_your_key_here
Content-Type: application/json

{
  "companyId": "your-company-id",
  "initiatedByUserId": "user-id-optional"
}

The subtask runs via the assigned agent's agent-task pipeline. Check mission status after to see output.

Auto-Execute the Entire Mission

Runs all subtasks sequentially with assigned agents:

POST https://app.thinkforce.ai/api/missions/{missionId}/auto-execute
X-TF-API-Key: tf_your_key_here
Content-Type: application/json

{
  "companyId": "your-company-id"
}

Publish Mission Report (PDF)

After a mission reaches completed, generate and publish a PDF report:

POST https://app.thinkforce.ai/api/publish-mission
X-TF-API-Key: tf_your_key_here
Content-Type: application/json

{
  "companyId": "your-company-id",
  "missionId": "mission-abc123",
  "userId": "user-id-optional"
}

Returns { publishedPdfUrl: "https://storage.googleapis.com/..." }.


Full Mission Workflow (JS Example)

const BASE = 'https://app.thinkforce.ai/api';
const HEADERS = { 'Content-Type': 'application/json', 'X-TF-API-Key': 'tf_your_key_here' };
const companyId = 'your-company-id';

// 1. Create mission
const { id: missionId } = await fetch(`${BASE}/missions`, {
  method: 'POST', headers: HEADERS,
  body: JSON.stringify({ companyId, title: 'Q2 Campaign', description: 'Research + content plan', priority: 'high' })
}).then(r => r.json());

// 2. Auto-decompose into subtasks
const { subtasks } = await fetch(`${BASE}/missions/${missionId}/decompose`, {
  method: 'POST', headers: HEADERS,
  body: JSON.stringify({ title: 'Q2 Campaign', description: 'Research + content plan' })
}).then(r => r.json());

// 3. Add each subtask (assign agents as needed)
for (const st of subtasks) {
  await fetch(`${BASE}/missions/${missionId}/subtasks`, {
    method: 'POST', headers: HEADERS,
    body: JSON.stringify({ companyId, ...st, assignedAgentId: 'agent-id' })
  });
}

// 4. Auto-execute all subtasks
await fetch(`${BASE}/missions/${missionId}/auto-execute`, {
  method: 'POST', headers: HEADERS,
  body: JSON.stringify({ companyId })
});

// 5. Mark complete + publish PDF
await fetch(`${BASE}/missions/${missionId}`, {
  method: 'PATCH', headers: HEADERS,
  body: JSON.stringify({ companyId, status: 'completed' })
});
const { publishedPdfUrl } = await fetch(`${BASE}/publish-mission`, {
  method: 'POST', headers: HEADERS,
  body: JSON.stringify({ companyId, missionId })
}).then(r => r.json());

Notes

  • CEO agent is always created first during onboarding and orchestrates your team automatically.
  • Model routing: If your account is linked to OpenAI Codex (ChatGPT Pro OAuth), tasks route through gpt-5.4 automatically.
  • Tools available: Google Search, Website Fetch, Memory Manager, Web Browser, and more — enabled per agent.
  • API key is per-company: One key covers all agents. Rotate anytime from Settings.

平台分布

OpenClaw

79.18%
按下载量换算1,272

安全审计

VirusTotal

通过

ClawScan

通过

Static analysis

通过

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills