← 返回 Skills 市场
lephix

YNote Clip

作者 lephix · GitHub ↗ · v1.8.1
cross-platform ⚠ suspicious
360
总下载
0
收藏
1
当前安装
1
版本数
在 OpenClaw 中安装
/install ynote-clip
功能描述
网页剪藏到有道云笔记。触发词:剪藏网页、保存网页、收藏网页、ynote clip、clip to ynote、网页摘录。
安全使用建议
Before installing, review and confirm these items: 1) The embedded DEFAULT_APIFY_TOKEN in twitter-apify.mjs — this allows the skill to use someone else's Apify account (privacy, billing, and abuse risk). Ask the author to remove the hard-coded token and require APIFY_API_TOKEN as an explicit env var. 2) Inspect the base64 payload in static/inject-sdk.fn.js (the atob string) — decode it locally to verify it contains only the intended collect SDK and not hidden exfiltration or remote endpoints. 3) Decide whether you are comfortable the skill asks you to source your shell rc (which exposes any exported secrets) each run; prefer setting YNOTE_API_KEY (and APIFY_API_TOKEN if used) explicitly in the agent's environment rather than relying on sourcing rc files. 4) If you will use the Twitter flow, prefer supplying your own APIFY_API_TOKEN or disable that flow; running with the default token may send scraped content to that third-party account. 5) Consider running the skill in a restricted/sandboxed environment first, and monitor outgoing network calls (especially to apify.com and the MCP endpoint) to ensure they match expectations. If you cannot verify the injected SDK and the embedded token, treat the skill as risky and avoid installing it.
功能分析
Type: OpenClaw Skill Name: ynote-clip Version: 1.8.1 The skill bundle contains several high-risk patterns and poor security practices. Most notably, SKILL.md instructs the AI agent to 'source ~/.zshrc' or other shell configuration files to load environment variables, which is a dangerous practice that can lead to the leakage of sensitive credentials or the execution of arbitrary shell aliases/functions. Additionally, twitter-apify.mjs contains a hardcoded Apify API token (apify_api_vsVgnrJKGDCfhfTil8FpBaMaM4vexW1TZocw), and inject-sdk.fn.js uses a base64-encoded string to deliver a large minified JavaScript payload for browser execution. While these features appear intended to facilitate web scraping and clipping to Youdao Note, the combination of environment manipulation, hardcoded credentials, and obfuscated payloads warrants a suspicious classification.
能力评估
Purpose & Capability
Name/description (网页剪藏到有道云笔记) aligns with the scripts and tools: node scripts, a browser injector, and MCP calls to create notes. Required binaries (node, curl, jq, dig) and browser tooling are plausible for the stated functionality. However, the code also integrates Apify (twitter-apify.mjs) and includes a DEFAULT_APIFY_TOKEN hard-coded in the file — this external service is only needed for Twitter/X special handling and the SKILL.md does not declare APIFY_API_TOKEN in requires.env, which is an omission and worth questioning.
Instruction Scope
SKILL.md directs the agent to run multiple shell scripts, write/read /tmp/ynote-clip-data.json, source the user's shell rc (e.g., source ~/.zshrc) to obtain environment variables, call web_fetch, and silently suppress certain errors (e.g., cron check). It intentionally keeps large bodyHtml out of the agent context (good), but it instructs silent failure paths and sourcing user shell files which may expose more of the user's environment than necessary. The twitter path calls Apify; SKILL.md mentions Apify usage but fails to declare APIFY_API_TOKEN as a required env var.
Install Mechanism
There is no install spec (instruction-only + bundled scripts). No network download/install steps are executed by an installer, so install-risk is low. The skill will execute local scripts and Node.js files that are bundled with it.
Credentials
Declared primary credential YNOTE_API_KEY is appropriate. However: (1) twitter-apify.mjs embeds a DEFAULT_APIFY_TOKEN in source code, enabling calls to Apify without the user's token — this is a high-risk hard-coded credential (can be abused or cause third-party charges). (2) SKILL.md and requires.env do not list APIFY_API_TOKEN even though the Twitter flow will use it if present — omission reduces transparency. The scripts also ask users to source shell rc files each run, which can leak any exported secrets present there to the child processes.
Persistence & Privilege
always is false; the skill does not request forced/global inclusion. It does invoke system tools and reads/writes /tmp files and uses openclaw cron/list commands, but it does not modify other skills or system-wide configs. Autonomous invocation is allowed (platform default) and not by itself a concern here.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install ynote-clip
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /ynote-clip 触发
  4. 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
v1.8.1
有道云笔记收藏网页
元数据
Slug ynote-clip
版本 1.8.1
许可证
累计安装 1
当前安装数 1
历史版本数 1
常见问题

YNote Clip 是什么?

网页剪藏到有道云笔记。触发词:剪藏网页、保存网页、收藏网页、ynote clip、clip to ynote、网页摘录。 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 360 次。

如何安装 YNote Clip?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install ynote-clip」即可一键安装,无需额外配置。

YNote Clip 是免费的吗?

是的,YNote Clip 完全免费(开源免费),可自由下载、安装和使用。

YNote Clip 支持哪些平台?

YNote Clip 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 YNote Clip?

由 lephix(@lephix)开发并维护,当前版本 v1.8.1。

💬 留言讨论