← 返回 Skills 市场
2915
总下载
7
收藏
16
当前安装
12
版本数
在 OpenClaw 中安装
/install think-cog
功能描述
AI brainstorming and strategic thinking powered by CellCog. Reasoning, problem-solving, ideation, strategic planning — then execution across every modality:...
安全使用建议
This skill appears to be what it says: a client wrapper to talk to CellCog for iterative thinking. Before installing/using:
- Only provide a CELLCOG_API_KEY you trust and scope it with least privilege if possible; treat it like any API secret.
- Install the 'cellcog' Python package only from a trusted source (official PyPI/project site) to avoid typosquatting or malicious packages — the skill lists 'cellcog' as a dependency but has no install step.
- Review CellCog's privacy/security documentation on cellcog.ai to understand what data is sent and retained by their service before sending sensitive data.
- If you need stronger assurance, ask the publisher for an explicit install guide or a package origin (PyPI name or GitHub repo) and verify the package fingerprint/source before installing.
功能分析
Type: OpenClaw Skill
Name: think-cog
Version: 1.0.11
The 'think-cog' skill bundle is a documentation-focused wrapper for the CellCog AI service, providing instructions and code examples for collaborative brainstorming. It requires the 'cellcog' Python dependency and a standard API key (CELLCOG_API_KEY), with no evidence of malicious execution, data exfiltration, or harmful prompt injection in SKILL.md or _meta.json.
能力标签
能力评估
Purpose & Capability
Name/description advertise interactive reasoning and execution via CellCog. The skill only requires python3 and CELLCOG_API_KEY and references the cellcog SDK — these are consistent with a client SDK-based integration for CellCog services.
Instruction Scope
SKILL.md contains usage examples for the CellCog SDK (create_chat, send_message) and instructs back-and-forth conversation. It does not instruct reading unrelated system files, harvesting unrelated environment variables, or contacting unexpected endpoints. It does reference the broader cellcog SDK for file handling and modes, which is expected.
Install Mechanism
This is instruction-only (no install spec). SKILL.md lists a dependency 'cellcog' but does not provide an install step — the runtime depends on the environment already having the cellcog package. This is reasonable but means the runtime will fail unless the SDK is installed; verify and install the official package yourself rather than running untrusted install commands.
Credentials
Only CELLCOG_API_KEY is required, which is appropriate for an SDK client that calls a remote CellCog API. There are no unrelated credentials or unusual environment/config path requests.
Persistence & Privilege
The skill is not always-enabled and does not request any elevated or persistent privileges. Autonomous invocation is allowed by default on the platform but that is expected for skills and not excessive here.
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 think-cog - 安装完成后，直接呼叫该 Skill 的名称或使用
/think-cog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
v1.0.11
- Added required Python binary (`python3`) and `CELLCOG_API_KEY` environment variable to skill metadata
- No changes to the skill's logic or usage instructions
- Documentation update for clearer dependency requirements
v1.0.10
- Updated the skill description for improved clarity and emphasis on AI-powered brainstorming, reasoning, and strategic planning.
- Revised SDK usage instructions for agents, clarifying agent compatibility and usage examples.
- Streamlined and clarified language throughout the documentation, focusing on accessibility and usability.
- No changes to functionality or interface; documentation only.
v1.0.9
**think-cog 1.0.9 Changelog**
- Improved getting started code example: now includes full initialization for the `CellCogClient` in Python.
- Expanded skill description to emphasize full-lifecycle reasoning and execution support.
- Minor copy-edits and streamlining of documentation language for clarity.
v1.0.8
- Updated skill description and "How to Use" section for improved clarity and conciseness.
- Added direct link to https://cellcog.ai for the complete SDK API reference.
- Simplified setup instructions and agent usage examples.
- Minor editorial improvements to sections for consistency and easier onboarding.
- No functional or code changes—documentation only.
v1.0.7
- Major rewrite of documentation for clarity and practical guidance
- Focus shifted to iterative, back-and-forth thinking and tangible output, not just conversation
- Added detailed use cases across technical, business, creative, and debugging scenarios
- Provided practical session tips and real-world examples to illustrate the skill’s value
- Clarified chat usage: favors collaborative agent mode for multi-turn problem-solving
- Comparative table added: distinguishes think-cog from fire-and-forget skills
v1.0.6
- Updated description and top summary to focus on collaborative, iterative problem-solving and clarify conversational nature.
- Streamlined documentation, emphasizing core usage patterns, prerequisites, and integration with CellCog.
- Added section listing CellCog's internal capabilities (reasoning engine, modality support, conversational threading).
- Clarified instructions for using send_message() and chat_mode="agent" for multi-turn dialogue.
- Condensed guidance on use cases, making it easier to understand appropriate scenarios for think-cog.
- Added a "Related Skills" section for easier discovery of complementary capabilities.
v1.0.5
**Changelog for think-cog v1.0.5**
- Added usage examples differentiating OpenClaw agents (fire-and-forget) and all other agents (blocking) to clarify session initiation steps.
- Updated prerequisite instructions to highlight `notify_session_key` for OpenClaw workflows.
- No functional or behavioral changes; documentation improved for installation and API usage clarity.
v1.0.4
- Updated the usage section to clarify that think-cog is conversational, not single-shot, and added an explicit code example for continuing a session.
- Improved integration guidance by referring users to the cellcog skill for the full SDK API reference.
- Minor edits for conciseness and clarity throughout the documentation.
v1.0.3
- Added operating system metadata (`os: [darwin, linux, windows]`) to skill description.
- Set `homepage` field to https://cellcog.ai for clearer documentation.
- No functional or behavioral changes to the skill logic.
v1.0.2
- Added author and dependencies fields to SKILL.md metadata.
- Minor formatting and clarity improvements throughout documentation.
- Updated prerequisites to reference the `cellcog` skill with improved formatting.
- No functional code or skill logic changes in this release.
v1.0.1
- Updated skill description and positioning to emphasize that think-cog enables iterative idea development, reasoning, and execution across multiple modalities.
- Reframed think-cog as an "Alfred who builds," highlighting the loop of Think → Do → Review → Repeat beyond just conversation.
- Stressed that CellCog can assist not only in brainstorming and problem-solving but also in producing real outputs (documents, data, prototypes) for hands-on iteration.
- Retained usage instructions, philosophy, and examples while clarifying that this skill is ideal for situations where both the problem and solution emerge through multi-step collaboration.
v1.0.0
- Initial release of think-cog: a collaborative thinking partner for brainstorming, reasoning, exploratory problem-solving, idea development, strategy, and decision support.
- Focuses on back-and-forth conversational sessions (not fire-and-forget) using CellCog’s agent chat mode.
- Designed for problems where both the problem and solution need to be discovered through dialogue.
- Use cases include architecture decisions, business strategy, creative direction, debugging, and structured decision making.
- Requires CellCog mothership skill for setup; see instructions for prerequisites and tips for effective sessions.
元数据
常见问题
Think Cog 是什么？
AI brainstorming and strategic thinking powered by CellCog. Reasoning, problem-solving, ideation, strategic planning — then execution across every modality:... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 2915 次。
如何安装 Think Cog？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install think-cog」即可一键安装，无需额外配置。
Think Cog 是免费的吗？
是的，Think Cog 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Think Cog 支持哪些平台？
Think Cog 跨平台运行，可在任意部署了 OpenClaw / Claude Code 的环境中使用（darwin, linux, windows）。
谁开发了 Think Cog？
由 CellCog（@nitishgargiitd）开发并维护，当前版本 v1.0.11。
推荐 Skills