跳到正文
cccskills
免费在 GitHub 公开

ask-agents

Send a specific question or task to the external agent or agents the user names and relay their responses. Use for an outside opinion or explicit request to ask Claude, Codex, or Gemini.

中文简介正在准备中。当前显示原文说明。

chrisblattman/claudeblattman星标 470(GitHub 470 + cccskills 0)
集成工具Gemini

实测评分 未实测

查看安装方法

包含的文件(5)

  • SKILL.md2.2 KB
  • references/host.md1.3 KB
  • references/package.json321 B
  • references/transport.md1.6 KB
  • scripts/run-external-agent.mjs30.8 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

Ask Agents

Read references/host.md and references/transport.md next to this skill. They specify the installed app's runner and its actual capabilities. Ask Agents is request-exact: it does not expand a question into a council or add unrequested providers. Council manages its own panel and may share transport machinery without invoking this conversational workflow.

Before an external call, apply the existing privacy and send-authorization checks. Do not transmit confidential records, personal contact data, participant data, private transcripts, or unpublished sensitive work through these starter adapters. Offer a local answer or a redacted request for approval. Respect explicit authorization already given for this task and service; do not ask again for an unchanged authorized action. Otherwise explain the service and content and obtain send authorization. Never bypass host permissions.

For independent opinions, send the same self-contained question and permitted evidence to separate fresh invocations. Add a concrete specialist question when useful. Do not let a later reviewer see earlier answers. Multiple requested reviewers from one service require multiple calls and distinct report files.

Use the bundled transport as documented. State its actual limits rather than claiming it is Chris's unrestricted balanced runtime. Missing executables or authentication are setup failures; never fabricate an external response. Retain the full reply and actual execution/access status. Identify the requested and observed model separately when available; never infer observed identity from the requested name.

For a standalone question, relay the answer with a brief comparison and useful next action. If a larger workflow called the runner for reports, save each unabridged response to that workflow's archive and return it without performing an extra synthesis. Preserve separate partial/additional replies and failures. Only remove temporary data after all required reports and status records have been saved.

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

council

免费

Develop or review a plan, draft, workflow, decision, or argument using independent council members and a fresh synthesis. Use for an explicit council request, including dictated counsel. Discussing councils alone is not an invocation.

中文简介正在准备中。当前显示原文说明。

chrisblattman/claudeblattman4702026年10月25日 更新

done

免费

Save a short HANDOFF.md so the next session can resume. Use when the user asks to wrap up, finish, or capture the session.

中文简介正在准备中。当前显示原文说明。

chrisblattman/claudeblattman4702026年10月25日 更新

kit-hello

免费

Verify the starter kit's installed skill files and version, and explain the next step. Use when the user runs kit-hello or asks whether installation worked.

中文简介正在准备中。当前显示原文说明。

chrisblattman/claudeblattman4702026年10月25日 更新

kit-setup

免费

Check optional outside-agent connections and help complete their setup. Use when the user asks to connect agents, test Ask Agents, or troubleshoot the starter kit.

中文简介正在准备中。当前显示原文说明。

chrisblattman/claudeblattman4702026年10月25日 更新

查看 chrisblattman 的全部技能

报告此技能的问题