医学量表检索技能

医学量表信息检索与标准化报告生成。通过浏览器自动化检索量表背景、版权、CDISC 映射、统计方法等信息，生成飞书云文档并保存到量表知识库。

What to check before installing/using this skill: - Credentials and permissions: The skill expects Feishu API access and browser automation privileges, but the registry entry lists no required environment variables or primary credential. Ask the publisher how Feishu tokens/credentials are provided and where they are stored; do not hand over global credentials without understanding scope and storage. - Local filesystem writes and shell execution: The runtime instructions instruct the agent to run shell commands (curl, ls) and save PDFs to a hard-coded path (/Users/wangyafei/Downloads/...). If you install this skill, change default paths to a safe, configurable folder and confirm the agent is allowed to write there. Prefer running in an isolated environment or sandbox during initial testing. - Forced downloads: The skill enforces mandatory retrieval of the 'official PDF' (download or screenshot). Ensure you have legal right to download/archive the PDF (copyright/commercial restrictions); otherwise the agent may attempt to access paywalled or restricted content. - Limit autonomous actions: The skill can be invoked by the agent autonomously (default). If you are uncomfortable with automated browsing + shell downloads, require user consent for network/file operations or disable autonomous invocation until you verify behavior. - Verify implementation details: Request clarification from the maintainer about how Feishu credentials are supplied, whether tokens are transient, and whether the skill will ever send data to external endpoints other than Feishu and the public sources named. Ask them to remove hard-coded user paths and to declare required env vars (Feishu token, any proxy settings) in the metadata. - Test in a sandbox first: Run the skill in a restricted test account and sandboxed filesystem, with minimized credentials, to observe exactly what it does before granting it access to production data/accounts. If the publisher cannot or will not clarify these points (where credentials come from, configurable save path, and why no env vars are declared), treat the skill as high-risk and avoid using it with sensitive accounts or systems.

Type: OpenClaw Skill Name: medical-scale-research Version: 1.0.0 The skill uses the `exec` tool to run `curl` commands for downloading files to a hardcoded local directory (`/Users/wangyafei/Downloads/scales/`), which presents a significant command injection risk if the URL or scale name inputs are manipulated. While the intent appears to be medical research and report generation, the use of shell execution for file management and the presence of environment-specific hardcoded paths are high-risk patterns (SKILL.md, README.md).