← 返回 Skills 市场
sealiu1997

ZSXQ Digest

作者 sealiu1997 · GitHub ↗ · v0.1.0
cross-platform ✓ 安全检测通过
272
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install zsxq-digest
功能描述
Summarize recent posts from followed Knowledge Planet circles into a daily digest prioritizing key updates using local tokens or browser relay for access.
安全使用建议
This skill is coherent with its stated goal, but it deals with highly sensitive local authentication artifacts (zsxq_access_token and browser cookies). Before installing or running: 1) Inspect the included scripts (capture_browser_cookies.js, auth_bootstrap.py, run_browser_bootstrap.py, etc.) to ensure you understand what they do. 2) Keep token files under a gitignored directory (state/session.token.json) and never paste tokens into public channels. 3) If you use browser bootstrap, only supply wsUrl values that point to a local/controlled DevTools endpoint (localhost); do not point to remote or unknown hosts. 4) Run the skill in an isolated environment or test host if you are unsure. 5) Prefer manual review of any captured cookie files before finalization, and revoke/rotate the token if you suspect it was mishandled. If you want stronger assurance, ask the author for explicit manifests of which cookies/fields are extracted and for code signatures or reproducible builds.
功能分析
Type: OpenClaw Skill Name: zsxq-digest Version: 0.1.0 The zsxq-digest skill bundle is a highly functional tool designed to summarize updates from the Knowledge Planet (ZSXQ) platform. While it possesses high-risk capabilities—specifically the use of Chrome DevTools Protocol (CDP) in scripts like 'capture_browser_cookies.js' and 'prepare_zsxq_qr_bootstrap.js' to extract session cookies—these actions are transparently documented and strictly aligned with the stated purpose of local authentication bootstrap. The bundle demonstrates strong security awareness through 'scripts/package_public_release.py', which explicitly scrubs sensitive state files before packaging, and 'SKILL.md' instructions that mandate local-only storage of credentials in gitignored paths. No evidence of intentional data exfiltration, unauthorized remote control, or malicious prompt injection was found.
能力评估
Purpose & Capability
The name/description ask for digesting Knowledge Planet content and the repository contains collectors, deduping, scoring, rendering, and browser-bootstrap scripts that match that purpose. The presence of cookie/cdp helpers, QR bootstrap, and token-file handling is expected for the declared access modes (local token, browser relay, fetch fallback).
Instruction Scope
SKILL.md explicitly instructs the agent to read/write local state files (e.g., state/session.token.json, state/auth-bootstrap.json), validate auth, and, when necessary, connect to a local browser CDP endpoint (wsUrl) to capture cookies and page snapshots. These instructions stay within the declared purpose, but they grant the skill access to sensitive local browser cookies and require the operator to supply/maintain local token files — a legitimate need for this use case but high-sensitivity in practice. The docs repeatedly warn not to publish tokens, which is appropriate.
Install Mechanism
There is no external install/download step and no third‑party package pulled at runtime; the repo includes local scripts that will be executed. This is lower-risk than remote downloads, but the skill package itself contains executable scripts that operate on local files and local CDP endpoints, so users should inspect them before running.
Credentials
The skill declares no required env vars or config paths, yet the runtime docs and scripts expect a local state directory (state/session.token.json, state/cursor.json, etc.) and optionally a local CDP wsUrl. That is coherent with its token-first design, but the manifest omission is a mild metadata inconsistency: the skill should declare expected config paths. Also, while no external credentials are requested, the scripts will capture/use sensitive cookies/tokens — this is proportional to the task but requires explicit user caution.
Persistence & Privilege
The skill does not request always:true and appears to only write its own local state files (session token, cursor, auth-bootstrap state). It does not modify other skills or system-wide agent settings per the provided files. Autonomous invocation is allowed by default (normal), so treat any automated runs as able to access local tokens if present.
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 zsxq-digest
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /zsxq-digest 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
v0.1.0
Initial public release: token-first ZSXQ stream digest with browser bootstrap/recovery and sanitized packaging.
元数据
Slug zsxq-digest
版本 0.1.0
许可证
累计安装 0
当前安装数 0
历史版本数 1
常见问题

ZSXQ Digest 是什么?

Summarize recent posts from followed Knowledge Planet circles into a daily digest prioritizing key updates using local tokens or browser relay for access. 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 272 次。

如何安装 ZSXQ Digest?

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

ZSXQ Digest 是免费的吗?

是的,ZSXQ Digest 完全免费(开源免费),可自由下载、安装和使用。

ZSXQ Digest 支持哪些平台?

ZSXQ Digest 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 ZSXQ Digest?

由 sealiu1997(@sealiu1997)开发并维护,当前版本 v0.1.0。

💬 留言讨论