← 返回 Skills 市场
zack-dev-cm

Using Agent Skills Router

作者 Zakhar Pashkin · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ 安全检测通过
31
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install using-agent-skills-router
功能描述
Route ambiguous agent work to the right skill, command style, and effort level. Use when a task mentions skills, agents, workflows, lifecycle commands, revie...
使用说明 (SKILL.md)

Using Agent Skills Router

Use this skill as a lightweight dispatcher. It decides which skill, workflow, or role should handle a request, then keeps the active path proportional to the task.

Routing Workflow

  1. Restate the user goal in one sentence.
  2. Classify the request:
    • tiny: answer or edit directly; avoid ceremony.
    • implementation: choose the most specific build or domain skill.
    • review: choose a review skill or a reviewer persona.
    • ship: choose a lifecycle or release skill.
    • unclear: gather repo context first, then decide.
  3. Select the smallest useful skill set. Prefer one primary skill and at most one support skill.
  4. State the route briefly before executing: Using \x3Cskill> because \x3Creason>.
  5. If no skill matches, continue with normal Codex behavior and say no specialized skill was needed.

Skill Selection Rules

  • Prefer exact domain skills over broad lifecycle skills.
  • Prefer repository-native tools, tests, and docs over generic checklists.
  • Use a lifecycle skill only when the user wants planning, implementation, verification, review, or shipping as a process.
  • Use a high-risk review skill when the task touches security controls, credentials, data deletion, public release, money, production systems, or broad permissions.
  • Avoid stacking skills that repeat the same gate. One clear review path is better than several shallow ones.

Output Shape

For a routing-only answer, return:

Route: \x3Cskill-or-direct>
Reason: \x3Cone sentence>
Next action: \x3Cwhat Codex should do now>

For execution tasks, keep the routing note short and then proceed. Do not create docs or task files unless the user asks or the chosen skill requires durable artifacts.

Anti-Patterns

  • Do not turn small fixes into a full spec-plan-build-review cycle.
  • Do not load long reference files just to decide whether a skill applies.
  • Do not role-play multiple agents when the runtime cannot actually isolate their work.
  • Do not ask the user which skill to use when local context makes the choice clear.
安全使用建议
This skill appears safe to install as a routing aid. It may influence which other skills the agent chooses for a task, so users should still review any subsequent skill’s permissions or actions before allowing high-impact work.
功能分析
Type: OpenClaw Skill Name: using-agent-skills-router Version: 1.0.0 The skill bundle is a meta-orchestrator designed to help an AI agent select the appropriate sub-skill or workflow based on task complexity and risk. It contains no executable code, network requests, or data exfiltration logic. The instructions in SKILL.md and references/routing-patterns.md actually promote security best practices by explicitly directing the agent to use high-risk review skills when tasks involve credentials, security controls, or production systems.
能力评估
Purpose & Capability
The stated purpose is to route ambiguous agent work to an appropriate skill or direct action, and the provided files consistently support that purpose.
Instruction Scope
The instructions are bounded toward proportional routing: choose the smallest useful skill set, prefer one primary skill, state the route briefly, and continue normally if no specialized skill is needed.
Install Mechanism
There is no install spec, no executable code, no package dependency, and no required binary; the skill is instruction-only.
Credentials
The metadata declares no credentials, environment variables, config paths, OS restrictions, or runtime capabilities, which is proportionate for a routing guide.
Persistence & Privilege
No persistence, background activity, privilege escalation, file mutation, or account access is requested; the skill even advises not creating durable artifacts unless asked or required by a chosen skill.
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 using-agent-skills-router
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /using-agent-skills-router 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
v1.0.0
Initial public release from https://github.com/zack-dev-cm/codex-clawhub-skill-pack/releases/tag/v1.0.0.
元数据
Slug using-agent-skills-router
版本 1.0.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

Using Agent Skills Router 是什么?

Route ambiguous agent work to the right skill, command style, and effort level. Use when a task mentions skills, agents, workflows, lifecycle commands, revie... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 31 次。

如何安装 Using Agent Skills Router?

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

Using Agent Skills Router 是免费的吗?

是的,Using Agent Skills Router 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

Using Agent Skills Router 支持哪些平台?

Using Agent Skills Router 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 Using Agent Skills Router?

由 Zakhar Pashkin(@zack-dev-cm)开发并维护,当前版本 v1.0.0。

💬 留言讨论