← 返回 Skills 市场
255
总下载
1
收藏
0
当前安装
2
版本数
在 OpenClaw 中安装
/install multi-thread-dialogue
功能描述
多线程对话架构：主代理前台响应，子代理后台处理任务，支持打断、状态处理、两阶段审查和任务监控 / Multi-thread dialogue architecture: main agent responds in foreground, sub-agents handle tasks in background...
安全使用建议
This skill appears internally consistent and is instruction-only. Before installing or applying its recommendations: 1) Back up your AGENTS.md and any agent-config files so you can revert changes. 2) Inspect the actual AGENTS.md edits it recommends and ensure they don't introduce unexpected triggers or external endpoints. 3) Create the memory file in a controlled location and review its contents/permissions (it will store task IDs and state). 4) If you plan to run 'npx skills add ...', inspect the package that command would fetch from the registry first (the skill bundle here is only docs). 5) Note the skill author/source is not strongly verified in registry metadata; prefer installing only from trusted sources. If you want, I can (a) produce a diff-ready AGENTS.md snippet you can review before applying, or (b) check a specific AGENTS.md file for unexpected content.
功能分析
Type: OpenClaw Skill
Name: multi-thread-dialogue
Version: 1.1.1
The skill provides a conceptual framework and instructional templates for implementing a multi-agent task management system within OpenClaw. It primarily consists of Markdown documentation (SKILL.md, README.md) and configuration files (skill.yaml, _meta.json) that guide the agent on how to route tasks, manage sub-agent states, and track progress using a local JSON file (memory/multi-thread-tasks.json). No malicious code, data exfiltration attempts, or harmful prompt injections were identified.
能力评估
Purpose & Capability
Name/description describe multi-thread dialogue and sub-agent orchestration; the skill only asks you to update AGENTS.md (agent behavior rules) and create a local task state file, which are coherent with the stated capability. No unrelated binaries, credentials, or config paths are requested.
Instruction Scope
Instructions are concrete and limited: add a 'Multi-Thread Dialogue Mode' section to AGENTS.md, create memory/multi-thread-tasks.json, and use provided prompt templates. This stays within the stated purpose, but it does instruct you to modify your agent config (AGENTS.md) and to create persistent task state on disk — both are powers that change agent behavior and persist state, so review those changes before applying.
Install Mechanism
There is no automated install spec or code to download. The README shows a suggested 'npx skills add' command, but the skill bundle itself is instruction-only. Because nothing is fetched or executed by the skill automatically, the install risk is low — however, if you run the npx command yourself it will fetch code from the registry, so inspect that package before running.
Credentials
The skill requests no environment variables, credentials, or config paths. The only persistent artifact is a local JSON file for task state (memory/multi-thread-tasks.json), which is proportionate to a task-management skill.
Persistence & Privilege
always:false and no autonomous privilege escalation. Still, the skill explicitly tells you to modify AGENTS.md (agent behavior rules) and create a persistent memory file. Modifying AGENTS.md can change global agent behavior — expected for an orchestration skill, but worth a deliberate review and backup before applying.
如何使用
- 确保已安装 OpenClaw（本地或 Docker 部署）
- 在对话框中输入安装命令：
/install multi-thread-dialogue - 安装完成后，直接呼叫该 Skill 的名称或使用
/multi-thread-dialogue触发 - 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
v1.1.1
- Added full bilingual (Chinese/English) documentation and instructions throughout SKILL.md for broader accessibility.
- No functional or logic changes; primary update is extended documentation for installation, usage, features, configuration, and file structure in both languages.
- Clarified that all core logic remains implemented via AGENTS.md, with this skill focusing on guidance and templates.
- No code or underlying logic modifications in this version.
v1.1.0
Multi-Thread Dialogue v1.1.0
- Introduces a multi-threaded dialogue architecture, enabling the system to handle multiple tasks in parallel using sub-agents.
- Supports task interruption, state management with four clear statuses (DONE, DONE_WITH_CONCERNS, NEEDS_CONTEXT, BLOCKED), and monitors for long-running or stalled tasks.
- Models can be selected dynamically based on task complexity; includes specific rules for splitting simple vs. long tasks, as well as handling large context.
- Implements an optional two-phase review for improved quality and provides direct user controls such as task numbering and interruption commands.
- Configuration guidance and prompt templates are included for easy setup in AGENTS.md, along with default monitoring and concurrency limits.
元数据
常见问题
领导模块 是什么？
多线程对话架构：主代理前台响应，子代理后台处理任务，支持打断、状态处理、两阶段审查和任务监控 / Multi-thread dialogue architecture: main agent responds in foreground, sub-agents handle tasks in background... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 255 次。
如何安装 领导模块？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install multi-thread-dialogue」即可一键安装，无需额外配置。
领导模块 是免费的吗？
是的，领导模块 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
领导模块 支持哪些平台？
领导模块 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 领导模块？
由 philonis（@philonis）开发并维护，当前版本 v1.1.1。
推荐 Skills