太湖云水务知识库

水务设备知识库 AI 助手，提供设备查询、选型建议、参数对比等功能

Before installing: 1) Clarify required credentials — the registry claims none but SKILL.md and scripts require OPENAI_API_KEY and TAVILY_API_KEY (and they will call api.tavily.com and an embeddings endpoint). Only provide keys you trust and scope their permissions. 2) Note the skill uses absolute Windows D:\ paths — verify these map to safe, intended directories on your system; it is Windows-centric and not portable. 3) The audit logger records full user input and other fields to JSONL logs in the project folder; if you expect users to supply personal contact info (the skill asks for phone/company/contactor), consider redacting or limiting log retention and access. 4) The Tavily updater and embedding client perform outbound network requests; confirm you trust those external services and review privacy implications for sending documents/queries. 5) Because the registry metadata didn’t declare the env vars the code uses, ask the publisher to correct metadata or provide an explanation. 6) If you plan to run the scripts, inspect and test them in an isolated environment, replace placeholder API keys, and configure strict file permissions for the log and vector_store directories.

Type: OpenClaw Skill Name: water-knowledge-assistant Version: 1.0.0 The bundle is a legitimate RAG-based (Retrieval-Augmented Generation) assistant designed for water equipment knowledge management. It includes well-structured scripts for audit logging (audit-logger.py), local file synchronization (ingest-from-e-drive.py), and external news updates via the Tavily API (tavily-daily-update.py). The requested file system permissions are restricted to specific local workspace directories (D:\code\openclaw_lakeskill\), and the network access is limited to the stated purpose of updating the knowledge base. No evidence of malicious intent, data exfiltration, or harmful prompt injection was found.