Token导航 LogoToken导航TokenDH.com
待分类权限需确认github未标认证来源可访问许可证需确认审计通过

synthesis-tree-of-thought综合思想树

Agent Skill

synthesis-tree-of-thought 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。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。

总安装

294

周安装

12

GitHub Stars

3

下载量

95
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

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:synthesis-tree-of-thought(综合思想树)
来源仓库:https://github.com/rajivpant/synthesis-skills
仓库路径:skills/synthesis-tree-of-thought
安装命令:
npx skills add https://github.com/rajivpant/synthesis-skills --skill synthesis-tree-of-thought
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 npx skills 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。

skills.shnpx skills
npx skills add https://github.com/rajivpant/synthesis-skills --skill synthesis-tree-of-thought

简介

synthesis-tree-of-thought 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中围绕仓库状态、代码变更或协作事项进行整理时使用。

  • 适用于复杂问题分解与协作流程管理,可结合原始 README 验证功能细节。
  • 通过 npx skills add 命令从指定 GitHub 仓库安装,需确认权限范围和维护状态。
  • 安装前建议确认是否会触发联网、命令执行或文件读写等操作。
  • 适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

SKILL.md

Tree of Thought

A reasoning technique that simulates multiple domain experts brainstorming step by step, critiquing each other's work, backtracking on flaws, and converging on a well-vetted conclusion.

How it works

  1. Multiple experts each write one step of their thinking, then share it with the group.
  2. Each expert critiques their own reasoning and the reasoning of every other expert.
  3. Experts validate their assertions against their domain knowledge.
  4. All experts proceed to the next step, incorporating feedback from the group.
  5. If any expert identifies a flaw in their logic, they backtrack to where the flaw occurred and restart that line of reasoning.
  6. If any expert realizes they are wrong, they acknowledge it and begin a new train of thought.
  7. Each expert assigns a confidence level (likelihood of correctness) to their current assertion.
  8. The process continues until the experts converge on a conclusion.

Template 1: General-purpose

Use this template for any question or problem that benefits from multi-perspective reasoning.


Imagine three different experts are answering this question. They will brainstorm the answer step by step, reasoning carefully and taking all facts into consideration. All experts will write down one step of their thinking, then share it with the group. They will each critique their response and all the responses of others. They will check their answer based on established knowledge and sound reasoning. Then all experts will go on to the next step and write down this step of their thinking. They will keep going through steps until they reach their conclusion, taking into account the thoughts of the other experts. If at any time they realize that there is a flaw in their logic, they will backtrack to where that flaw occurred. If any expert realizes they are wrong at any point, they acknowledge this and start another train of thought. Each expert will assign a likelihood of their current assertion being correct. Continue until the experts agree on the single most likely answer.

The question is: [INSERT QUESTION]


Template 2: Domain-expert variant

Use this template when the problem requires specific domain expertise. Define the experts to match the problem domain.


Imagine three different deeply knowledgeable and experienced experts. One is an expert in [DOMAIN 1]. The second is an expert in [DOMAIN 2]. The third is an expert in [DOMAIN 3].

They are helping with the following task. They will brainstorm the answer step by step, reasoning carefully and taking all facts into consideration. All experts will write down a first draft of their thinking, then share it with the group. They will each critique their response and all the responses of others. They will check their drafts based on their expertise in their respective fields. Then all experts will go on to the next step and write down the next draft of their thinking. They will keep going through steps until they reach their conclusion, taking into account the thoughts of the other experts. If at any time they realize that there is a flaw in their logic, they will backtrack to where that flaw occurred. If any expert realizes they are wrong at any point, they acknowledge this and start another train of thought. Each expert will assign a likelihood of their current assertion being correct. Continue until the experts agree on a comprehensive result.

The task is: [INSERT TASK]


Guidance for choosing experts

  • Match expert domains to the problem. A technical question needs technical experts; a business strategy question needs business, economics, and industry experts.
  • Three experts is the standard. Use four when the problem spans four genuinely distinct domains.
  • Name specific expertise areas rather than generic titles. "Expert in distributed systems and database internals" is better than "technology expert."
  • Include at least one expert whose domain creates productive tension with the others (e.g., a security expert alongside a UX expert, or an economist alongside an engineer).

When to use tree of thought

  • Complex problems with no obvious single correct answer
  • Decisions where trade-offs matter and different perspectives reveal different trade-offs
  • Questions where domain expertise from multiple fields is required
  • Situations where premature convergence on a wrong answer is a risk

When NOT to use tree of thought

  • Simple factual lookups
  • Tasks with a single well-defined correct answer
  • When speed matters more than depth

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

35.96%
按下载量换算34

Claude

29.53%
按下载量换算28

Cursor

19.2%
按下载量换算18

Gemini CLI

8.52%
按下载量换算8

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

权限需确认

当前来源未能明确判断权限范围,默认进入异常复核队列。

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。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills