← 返回 Skills 市场
Task Dispatch
作者
CCCaptain0129
· GitHub ↗
· v1.1.0
· MIT-0
148
总下载
0
收藏
1
当前安装
2
版本数
在 OpenClaw 中安装
/install task-dispatch
功能描述
Task scheduling and dispatching for task boards. Use when setting up periodic task dispatch, checking for dispatchable tasks, creating subagents to execute t...
安全使用建议
Before installing or following these instructions: (1) review the ClawBoard GitHub repository being cloned — do not run unreviewed code as root; (2) confirm why TASKBOARD_API_URL and TASKBOARD_ACCESS_TOKEN are needed (they are required at runtime) and store tokens securely (avoid placing secrets in world‑readable files); (3) run the deployment in an isolated test environment (container/VM) first so any third‑party code cannot access sensitive host data; (4) limit subagent permissions and concurrency (maxConcurrentTasks=1) and set sensible timeouts to avoid runaway continuous dispatch; (5) if you expect the skill to declare required env vars, ask the author to update metadata so required credentials are explicit. If you cannot audit the cloned repo, do not run the install/start/token commands on a production machine.
功能分析
Type: OpenClaw Skill
Name: task-dispatch
Version: 1.1.0
The skill automates the deployment and orchestration of a task management system called 'ClawBoard'. It performs high-risk operations including cloning an external GitHub repository (github.com/CCCaptain0129/ClawBoard.git), executing local scripts (./clawboard install), and programmatically extracting and moving access tokens between environment files. While these actions are consistent with the stated purpose of setting up a task-dispatching service, the automated installation of external code and credential manipulation represent significant security risks that require manual verification of the external repository.
能力评估
Purpose & Capability
The skill's name and description (task scheduling/dispatch) align with the files and instructions. However, the package metadata declares no required environment variables or credentials while SKILL.md and README repeatedly require TASKBOARD_API_URL and TASKBOARD_ACCESS_TOKEN and tell the agent to write/read .env files. That mismatch (declared requirements = none vs instructions requiring tokens) is an incoherence a user should be aware of.
Instruction Scope
SKILL.md instructs running commands that clone and run a third‑party project (git clone https://github.com/CCCaptain0129/ClawBoard.git and ./clawboard install/start/token), reading ~/ClawBoard/.env, writing a .env into the agent workspace (~/.openclaw/workspace-<name>/.env), and spawning subagents in a continuous dispatch loop. These actions include filesystem reads/writes and execution of external code beyond simple API calls — broader scope than a purely declarative dispatcher and potentially dangerous if the cloned code is untrusted.
Install Mechanism
There is no formal install spec (instruction-only), which is low-risk by itself, but the instructions explicitly direct cloning and executing a third‑party GitHub repository. That is a moderate risk because it causes arbitrary remote code (the ClawBoard repo) to be installed/run on the host; the skill itself does not bundle that code but relies on it.
Credentials
The skill uses and instructs creation of TASKBOARD_API_URL and TASKBOARD_ACCESS_TOKEN and expects tokens to be stored in filesystem .env files. Yet the declared required env vars list is empty. The token handling (writing token into agent workspace .env, reading token from ~/ClawBoard/.env) involves sensitive credentials and file paths; requesting/storing these credentials without declaring them is disproportionate and inconsistent.
Persistence & Privilege
The skill does not set always:true and does not claim to modify other skills or system settings. However it encourages continuous dispatch loops and spawning subagents until no tasks remain — combined with autonomous invocation (platform default) this can create a long‑running process and many subagent executions. This is not automatically disallowed, but it increases blast radius if the deployed ClawBoard or subagents are untrusted.
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 task-dispatch - 安装完成后，直接呼叫该 Skill 的名称或使用
/task-dispatch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
v1.1.0
新增 ClawBoard 部署引导、环境检测、服务启动、配置向导功能
v1.0.0
task-dispatch v1.0.0
- Initial release: automated scheduling and dispatching for task boards (e.g., ClawBoard).
- Dispatches tasks continuously in a loop until no dispatchable tasks remain.
- Spawns subagents to execute tasks, waits for completion, verifies results, and updates task status.
- Includes detailed rules for task selection, status transitions, deliverable and criteria verification, and failure handling.
- Provides API usage, configuration pointers, and output templates for integration.
元数据
常见问题
Task Dispatch 是什么？
Task scheduling and dispatching for task boards. Use when setting up periodic task dispatch, checking for dispatchable tasks, creating subagents to execute t... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件，目前累计下载 148 次。
如何安装 Task Dispatch？
在 OpenClaw 或 Claude Code 对话框中运行命令「/install task-dispatch」即可一键安装，无需额外配置。
Task Dispatch 是免费的吗？
是的，Task Dispatch 完全免费，采用 MIT-0 许可证，可自由下载、安装和使用。
Task Dispatch 支持哪些平台？
Task Dispatch 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 Task Dispatch？
由 CCCaptain0129（@cccaptain0129）开发并维护，当前版本 v1.1.0。
推荐 Skills