Token导航 LogoToken导航TokenDH.com
研究检索需要联网github未标认证来源可访问许可证需确认审计提醒

relay-operator中继操作员

Agent Skill

relay-operator 用于查找、检索和筛选相关信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要根据关键词、任务场景或来源线索快速定位候选结果时使用。可结合来源仓库、安装命令和原始 README 继续核验具体用法。安装前建议确认权限范围、维护状态,以及是否会触发联网、命令执行或文件读写。

总安装

216

周安装

9

GitHub Stars

1

下载量

72
CodexClaudeCursorGemini CLI

安装说明

本站只整理中文说明和来源信息,不托管安装包,也不代用户安装。

GitHub

来源数

2

许可证

unknown

最后核验

2026-05-01

来源状态

来源可访问

安装方式

通过对话安装

复制提示词发给支持本地命令或 Skills 的 AI 助手,先确认命令和权限,再让它执行。

请帮我安装这个 Agent Skill:relay-operator(中继操作员)
来源仓库:https://github.com/eddiearc/relay
仓库路径:skills/relay-operator
安装命令:
npx skills add https://github.com/eddiearc/relay --skill relay-operator
安装前请先检查当前环境是否支持对应 CLI,并向我确认将要执行的命令、安装目录、联网范围和文件读写权限;确认后再执行。

命令行安装

复制命令到本机终端执行。该命令会通过 npx skills 从第三方来源获取 Skill;本站只展示命令,不托管安装包,也不自动执行。

skills.shnpx skills
npx skills add https://github.com/eddiearc/relay --skill relay-operator

简介

relay-operator 用于查找、检索和筛选相关信息。

  • 适合在 Codex、Claude、Cursor、Gemini CLI 中根据关键词或任务场景快速定位候选结果。
  • 通过 npx skills add 命令从指定 GitHub 仓库安装并使用。
  • 安装前需确认权限范围、维护状态,注意是否触发联网、命令执行或文件读写操作。
  • 适用宿主包括 Codex、Claude、Cursor、Gemini CLI,接入前应确认版本、权限和运行环境要求。

SKILL.md

Relay Operator

This skill is only the agent adapter layer.

Treat relay-operator as the current best-practice path, not as a frozen or exclusive workflow. If you find a path that is clearer, safer, or more effective than the current one, prefer improving Relay itself and sending a GitHub PR to refine this skill and the CLI guidance.

Before doing Relay work on first use:

command -v relay >/dev/null 2>&1 || npm install -g @eddiearc/relay || go install github.com/eddiearc/relay/cmd/relay@latest
relay upgrade --check
relay help

For concrete operations, use the Relay CLI as the source of truth:

  • relay help
  • relay help watch
  • relay help pipeline
  • relay help issue
  • relay pipeline list
  • relay pipeline show <name>
  • relay pipeline template
  • relay issue template
  • relay issue add --pipeline <name> --goal "..." --description "..."
  • relay watch -issue <id>
  • relay help upgrade

Agent runner selection:

  • relay pipeline add auto-detects available runners (codex, claude) from PATH when --agent-runner is not specified
  • if you are running inside Claude Code, suggest --agent-runner claude to the user
  • if you are running inside Codex CLI, suggest --agent-runner codex to the user
  • if the user does not specify, the CLI will detect and select the first available runner automatically

Rules:

  • do not restate Relay workflows from memory when CLI help already covers them
  • inspect the current repository before choosing a pipeline
  • use relay pipeline list and relay pipeline show to inspect candidates
  • if one pipeline is clearly the right fit for the current repo and task, select it directly
  • if multiple pipelines are plausible or the fit is unclear, present the candidates and ask the user to choose
  • if no saved pipeline is clearly suitable, start from relay pipeline template, write a repository-specific pipeline, and import it
  • when selecting or creating a pipeline, do not silently proceed; first give the user a short directional summary and ask whether it sounds right
  • summarize the pipeline at the intent level instead of dumping full prompt text by default; cover:

- what planning should focus on for this repo and task - what coding should focus on, including important repo constraints and reusable project assets - how the work should be verified - whether reusable E2E or unit-test coverage already exists - if the task is a good E2E candidate and reusable E2E coverage is missing, suggest adding a reusable relay-e2e scenario or another suitable verification skill - if unit tests are missing, call that out more strongly and recommend adding them

  • before creating or editing an issue, say plainly where the user's task definition is weak or risky; if the goal, scope, non-goals, or verification path are underspecified, fix that with the user instead of silently guessing
  • once the pipeline and issue direction looks good, explain the key directional decisions in terms of the chosen pipeline config and prompt intent, then ask the user to confirm that the direction sounds right before proceeding
  • treat project-level verification as harness-critical, not optional
  • use the strongest realistic verification path as the default for meaningful behavior changes, but allow narrower verification when the task genuinely does not justify the heavier path; when making that exception, say so explicitly and explain why
  • use relay help pipeline and relay help issue for the concrete verification defaults by project shape
  • if the repo lacks the right verification layer, call that out explicitly and recommend the missing script, test suite, or skill
  • rewrite the task into a Relay-ready issue before creating it
  • if the issue is still too vague, ask the user several focused questions in one turn before creating it; those questions should cover:

- the target end result - scope limits and explicit non-goals - how completion should be verified

  • use relay watch for user-facing monitoring; keep relay serve as the executor
  • do not duplicate templates or deep operational guidance in this skill
  • if CLI help is insufficient, improve the CLI help instead of expanding this file
  • if the bundled skill itself should be refreshed, use: npx skills add https://github.com/eddiearc/relay --skill relay-operator -g -y

适合场景

01

用户想查找某类 Agent Skill 时

02

需要根据任务场景推荐可安装能力包时

03

需要对比不同来源的安装命令和来源信息时

能力概览

能力 1

按任务关键词查找相关 Skills

能力 2

展示可复制的安装命令

能力 3

保留来源站点、仓库和原始说明,方便继续核验

能力 4

展示第三方安全扫描或审计结果

安装后应在对应宿主中按原始 README 的触发条件使用;具体调用方式请以来源页面和 README 为准。

平台分布

Codex

37.32%
按下载量换算27

Claude

29.65%
按下载量换算21

Cursor

18.96%
按下载量换算14

Gemini CLI

8.92%
按下载量换算6

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

安装前确认

本站仅展示第三方公开信息,不托管安装包,不提供自动安装或运行环境。安装前应自行审查源码、依赖和命令行为。来源安全扫描存在 warning/failed 结果,不能写成本站确认安全。当前只有一个来源,正式发布前建议补源仓库或其他目录站核验。

来源信息

继续浏览同类 Skills