← 返回 Skills 市场
DeepLink Agentic
作者
ShirleyDDDD
· GitHub ↗
· v1.1.0
· MIT-0
135
总下载
0
收藏
0
当前安装
3
版本数
在 OpenClaw 中安装
/install realestate-deep-research
功能描述
深度智联 Agentic 智能体 Skill。让用户告别碎片化查询，由 AI 智能体统筹复杂的房地产行业研究任务。支持房地产市场分析、土地研判、房地产企业分析、房地产项目案例分析、房地产项目设计建议、政策解读、物业行业资讯（日/周/月报）、物业行业招投标监测等房地产行业研究任务的创建、进度查询、成果获取与迭代优化...
安全使用建议
This skill appears to do what it says: it forwards user queries and any user-uploaded files to agentic.dichanai.com using AGENTIC_TOKEN. Before installing, confirm you trust agentic.dichanai.com. Avoid uploading highly sensitive data unless you understand that it will be transmitted to that service. Be aware the token-renewal step prints a NEW_TOKEN value to the terminal — ensure your agent or workflow will not accidentally share that output with others or with the user. Prefer using a least-privilege or short-lived AGENTIC_TOKEN if possible, and rotate the token if you suspect it was exposed. Finally, note the update instructions which can replace local skill files from the remote host — keep that trust boundary in mind because compromised updates could change behavior.
功能分析
Type: OpenClaw Skill
Name: realestate-deep-research
Version: 1.1.0
The skill bundle is a functional integration for the 'Agentic' real estate research platform (agentic.dichanai.com). The primary component is a Python CLI script (agentic.py) that interacts with a documented API to create research tasks, manage files, and handle token lifecycle. While the script includes a token renewal feature that prints new credentials to stdout (potentially risky if the agent leaks terminal output), the SKILL.md instructions explicitly warn the AI agent to keep this data private. The code is well-structured, lacks obfuscation, and its behaviors (file uploads, network requests to the primary domain) are consistent with its stated purpose of performing deep industry research.
能力评估
Purpose & Capability
Name/description match implementation: the skill's CLI and SKILL.md talk to agentic.dichanai.com for creating/monitoring research tasks. Required binary (python3), the requests dependency, and a single AGENTIC_TOKEN are all expected for this purpose.
Instruction Scope
Instructions explicitly instruct the agent to send user queries and any user-uploaded files to agentic.dichanai.com and to use the provided CLI. The SKILL.md requires the agent to inform the user that data will be sent to agentic.dichanai.com (good). It also directs the agent to upload files and to poll for HITL states. This behavior is coherent, but it means user data and uploads will be transmitted to the remote platform (declared in the doc).
Install Mechanism
Install spec only pulls in the 'requests' package (reasonable). The skill also instructs manual update via fetching a .skill file from the platform; that implies a remote update path (host: agentic.dichanai.com). This is expected for a hosted service but poses the usual update risk: if the remote update source is compromised, updated skill files could change behavior.
Credentials
Only AGENTIC_TOKEN is required and used as the primary credential, which is proportionate. Two caution points: (1) the token renewal flow prints NEW_TOKEN to stdout and asks the operator/agent to update AGENTIC_TOKEN — if the agent forwards terminal output to users, the token could be leaked; (2) token scope is unspecified — a broadly privileged token would increase risk. Both are operational considerations rather than inconsistencies.
Persistence & Privilege
always is false and the skill does not request system-wide privileges or alter other skills. Autonomous invocation is allowed (platform default) but not combined here with other suspicious privileges.
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 realestate-deep-research - 安装完成后，直接呼叫该 Skill 的名称或使用
/realestate-deep-research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
v1.1.0
Version 1.1.0
- 明确说明 `poll` 命令不会自动处理中途出现的 HITL（人类介入）状态，需人工判断并中断。
- 优化 HITL 轮询说明，强调需主动转交用户，避免遗漏。
- 对使用流程做了小幅调整，强化了任务状态和用户指导方面的细节说明。
- 其余文本与命令未发生重大变动。
v1.0.1
Version 1.1.0 introduces HITL（Human-in-the-Loop）支持及安全更新：
- 新增 HITL 流程说明：任务遇到 hitl 状态时必须中止自动流程，主动向用户转述问题，获取意见后再提交，确保人机协作可控。
- 增加数据交互与隐私提示，明确文件传输范围及数据保护承诺。
- 安全提示：Token 自动续期不得在用户侧泄露，敏感信息只应内部更新。
- 标准工作流、命令用法说明全面更新，同步增强 demo 和环境/依赖说明。
- 新增依赖声明、元数据、环境变量与 requests 包自动安装指令，提升易用性。
v1.0.0
Initial release of realestate-deep-research Skill:
- Enables users to manage complex real estate research tasks through an AI agent tailored for the industry.
- Supports creation, progress tracking, result retrieval, and iterative refinement of tasks such as market analysis, land evaluation, enterprise research, policy interpretation, industry news, and bidding monitoring.
- Guides users in clarifying and structuring their research requirements before submission.
- All operations are performed via a unified script interface with detailed subcommands.
- Integrates token management and version update checking for secure and reliable access.
- Reminds users of point consumption on the platform before creating tasks.
- Provides comprehensive reference and best practices for task creation, file management, and status tracking.
元数据
常见问题
DeepLink Agentic 是什么？
深度智联 Agentic 智能体 Skill。让用户告别碎片化查询，由 AI 智能体统筹复杂的房地产行业研究任务。支持房地产市场分析、土地研判、房地产企业分析、房地产项目案例分析、房地产项目设计建议、政策解读、物业行业资讯（日/周/月报）、物业行业招投标监测等房地产行业研究任务的创建、进度查询、成果获取与迭代优化... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 135 次。
如何安装 DeepLink Agentic？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install realestate-deep-research」即可一键安装，无需额外配置。
DeepLink Agentic 是免费的吗？
是的，DeepLink Agentic 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
DeepLink Agentic 支持哪些平台？
DeepLink Agentic 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 DeepLink Agentic？
由 ShirleyDDDD（@shirleydddd）开发并维护，当前版本 v1.1.0。
推荐 Skills