zlm

定期拉取 ZLMediaKit 仓库，查看最新的 Issue、Pull Request，并提供源码深度分析功能。当用户要求查看 ZLMediaKit 动态、Issue、PR 或分析其源码时触发调用。

Security Scan

The skill's code and instructions align with its stated purpose (periodically pulling the ZLMediaKit repo, fetching Issues/PRs, and producing a local analysis report); nothing indicates hidden exfiltration or unrelated credential access, though there are a few minor inconsistencies to be aware of.

Purpose & Capability

The skill's declared top-level requirements (none) differ from the SKILL.md metadata that lists required tools (git, gh, curl, grep). The actual behavior (git clone/pull, GitHub API calls, local code analysis) legitimately needs git and network access and optionally gh/curl, so the tools listed in SKILL.md are appropriate; the mismatch between registry metadata and SKILL.md is an inconsistency but not malicious.

Instruction Scope

Runtime instructions and the included Python script only clone/pull the public GitHub repo, query GitHub's API for issues/PRs, search the local source tree and write a Markdown report. The skill does not execute code from the cloned repo or call unknown external endpoints beyond api.github.com. It does reference an optional GITHUB_TOKEN and a local workspace path (D:\.openclaw\workspace).

Install Mechanism

No install spec is provided (instruction-only with an included helper script). No remote installs or archive extraction occur. Risk is low because code writes only to the declared workspace path and uses standard CLI tools.

Credentials

The skill does not require secrets, but SKILL.md and the script reference optional environment variables (GITHUB_TOKEN to avoid rate limits and WORKSPACE_DIR to override the workspace). These env vars are proportional to the task, but they are not declared in the skill registry metadata — a minor transparency gap.

Persistence & Privilege

The skill is not always: true, does not request elevated privileges, and only writes files into its workspace. It does not modify other skills or system-wide settings.

Assessment

This skill appears to do what it says: clone/pull the public ZLMediaKit repository, query GitHub for issues/PRs, and generate a local Markdown report. Before installing or enabling autonomous runs, consider: 1) Confirm you are comfortable letting the agent write to the workspace path (default D:\.openclaw\workspace) or set WORKSPACE_DIR to a safe location. 2) The SKILL.md mentions git/gh/curl/grep — ensure those tools are present (gh is optional). 3) If you supply a GITHUB_TOKEN, treat it as a secret (it’s only used to raise GitHub API rate limits). 4) The registry metadata omits the listed tools and env usage — you may want the publisher to fix that metadata mismatch for clarity. 5) If you want extra caution, run the included python script manually to inspect its behavior before allowing the agent to invoke it autonomously.

ZLMediaKit Analyzer

该 Skill 用于定期跟踪和分析开源流媒体框架 ZLMediaKit 的最新动态，包括 Issue、Pull Request，并支持对其 C++11 源码进行深度分析。

环境要求

• 已安装 git (用于克隆和拉取代码)
• 已安装 GitHub CLI gh (用于获取 Issue 和 PR 列表)
• 可选：配置 GITHUB_TOKEN 环境变量以避免 API 速率限制

Workflow 工作流

当用户触发此 Skill 时，Agent 应按照以下步骤执行：

1. 检查并拉取最新代码 (Sync Code)

检查本地是否存在 ZLMediaKit 代码库（默认路径：D:\.openclaw\workspace\ZLMediaKit）。

• 如果不存在，使用 RunCommand 工具执行克隆：

  git clone https://github.com/ZLMediaKit/ZLMediaKit.git D:\.openclaw\workspace\ZLMediaKit
  

• 如果已存在，则进入目录并拉取最新代码：

  cd D:\.openclaw\workspace\ZLMediaKit
  git pull origin master
  

2. 获取最新动态 (Fetch Issues & PRs)

使用 RunCommand 工具调用 gh CLI 获取最新的 5 个 Issue 和 Pull Request：

• 查看最近的 Issues：

  gh issue list -R ZLMediaKit/ZLMediaKit --limit 5 --state all
  

• 查看最近的 Pull Requests：

  gh pr list -R ZLMediaKit/ZLMediaKit --limit 5 --state all
  

(注：如果没有 gh CLI，可以使用 curl -s https://api.github.com/repos/ZLMediaKit/ZLMediaKit/issues?per_page=5 替代)

3. 源码深度分析 (Code Analysis)

当用户要求分析特定模块或源码（如 RTSP/RTMP、WebRTC、网络 IO 模型等）时：

• 优先建议使用内置的 search agent (default_api:search) 或 SearchCodebase 工具，在 D:\.openclaw\workspace\ZLMediaKit 目录下搜索相关实现。
• 结合 ZLMediaKit 项目的特点进行解读：
  • 基于 C++11 开发，无裸指针，采用智能指针管理内存。
  • 使用 多路复用/多线程/异步网络 IO 模式。
  • 支持多种协议互转 (RTSP/RTMP/HLS/WebRTC 等)。
• 对关键的类（如 TcpServer, EventPoller, Session 等）或请求处理流程进行追踪，生成函数调用栈或架构解析。

4. 生成报告 (Generate Report)

根据用户需求，综合社区动态（最新的 Issue/PR 讨论）与源码分析结果，输出结构化、排版清晰的 Markdown 报告。报告中应包含代码文件的具体路径（如 src/Network/TcpServer.cpp）及核心逻辑解释。

计划任务 / 定期拉取说明

如需定期执行，可通过 OpenClaw 或系统级定时任务（如 Windows 任务计划程序、Cron）调用该 Skill 对应的执行脚本，自动触发 "拉取代码 -> 检查 PR -> 生成简报" 的工作流。