← 返回 Skills 市场
68
总下载
0
收藏
0
当前安装
4
版本数
在 OpenClaw 中安装
/install thinkdeep
功能描述
Structured reasoning protocol for Claude — forces step-by-step analysis, self-critique, and confidence scoring before answering. Reduces wrong answers and ha...
安全使用建议
This skill is instruction-only and appears coherent with its goal of producing more careful answers — it asks for no credentials and installs nothing. Before installing, consider: 1) chain-of-thought output can be more verbose and may expose internal reasoning in logs (avoid when that is sensitive); 2) if you prefer not to allow the agent to call skills autonomously, restrict or review autonomous invocation in your agent settings; 3) the README/SKILL naming/version inconsistencies are minor documentation issues but verify you install the intended skill name (thinkdeep) if following install commands. Overall there are no security red flags.
功能分析
Type: OpenClaw Skill
Name: thinkdeep
Version: 1.0.3
The 'thinkdeep' skill is a purely instructional reasoning protocol designed to improve AI accuracy through structured chain-of-thought analysis. It contains no executable code, scripts, or network requests. The instructions in SKILL.md and README.md are focused entirely on guiding the agent's internal logic (Restate, Explore, Critique, and Confidence Check) to reduce hallucinations, with no evidence of malicious prompt injection or data exfiltration.
能力评估
Purpose & Capability
The name, description, and SKILL.md all describe a structured reasoning protocol and the runtime instructions implement that protocol. Minor documentation inconsistencies exist (README and some install suggestions use 'deep-think' or 'deep think' while the skill slug is 'thinkdeep'; SKILL.md version 1.0.1 vs registry 1.0.3), but these are editorial and do not affect functionality or require extra permissions.
Instruction Scope
SKILL.md confines itself to prompting and output formatting: restate, explore, critique, select & reason, and confidence scoring. It does not instruct reading files, accessing environment variables, or contacting external endpoints beyond recommending installing other skills.
Install Mechanism
No install spec and no code files — instruction-only skill. Nothing is written to disk or fetched at install time.
Credentials
The skill declares no required environment variables, credentials, or config paths. There are no disproportionate secret or credential requests.
Persistence & Privilege
always:false and user-invocable:true (defaults). The skill may be invoked autonomously by agents (platform default), but it does not request elevated persistence or modify other skills or system settings.
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 thinkdeep - 安装完成后，直接呼叫该 Skill 的名称或使用
/thinkdeep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
v1.0.3
- Version bump from 1.0.1 to 1.0.3 with no file changes detected.
- No updates made to SKILL.md or other files.
- Functionality and documentation remain unchanged from prior version.
v1.0.2
No changes detected in this release.
- Version number remains at 1.0.1 in SKILL.md despite version label 1.0.2.
- No file changes or content updates identified.
v1.0.1
- Added author and homepage information to SKILL.md
- Expanded and refined the keywords list for improved discoverability
- Updated branding and terminology from "Deep Think" to "Thinkdeep" throughout
- Enhanced description to clarify the skill's purpose and benefits
- Adjusted installation instructions to match updated naming (thinkdeep)
v1.0.0
- Initial release of thinkdeep 1.0.0.
- Introduces a structured reasoning protocol for Claude to ensure step-by-step analysis, self-critique, and confidence scoring before answering complex questions.
- Reduces risk of wrong answers and hallucinations, especially for technical or high-stakes queries.
- Protocol includes restating the problem, exploring multiple approaches, critiquing each, selecting the best, and rating confidence.
- Guides on when to activate, output format, and anti-patterns to avoid.
元数据
常见问题
Thinkdeep 是什么？
Structured reasoning protocol for Claude — forces step-by-step analysis, self-critique, and confidence scoring before answering. Reduces wrong answers and ha... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 68 次。
如何安装 Thinkdeep？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install thinkdeep」即可一键安装，无需额外配置。
Thinkdeep 是免费的吗？
是的，Thinkdeep 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Thinkdeep 支持哪些平台？
Thinkdeep 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 Thinkdeep？
由 jiajiaoy（@jiajiaoy）开发并维护，当前版本 v1.0.3。
推荐 Skills