Token导航 LogoToken导航

Agent Skill · 研究检索

plan(计划)

plan 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。

plan 是研究检索类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为调用外部服务;许可证信息还不明确,第三方扫描已通过,按安装量在研究检索类的 73,675 个 Skill 里排第 64,068。

总安装 306周安装 13GitHub Stars 公开资料未说明下载 107核验 2026-05-01

安装说明

本站只整理中文说明,不托管安装包,也不代用户安装。
GitHub
许可证
unknown

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:plan(计划)
来源仓库:https://github.com/anortham/goldfish
仓库路径:skills/plan
安装命令:
npx skills add https://github.com/anortham/goldfish --skill plan
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/anortham/goldfish --skill plan

这个 Skill 是做什么的

研究检索类 Skill 主要解决「找资料、核对事实、把来源整理清楚」的问题,通常会让助手先检索再总结。

plan 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。创建持久化战略计划,确保多会话间方向一致。

安装命令:npx skills add https://github.com/anortham/goldfish --skill plan。需确认权限范围、维护状态及是否触发联网或文件读写。

来源仓库内路径:skills/plan

上手步骤

4 步
  1. 1

    确认自己的宿主工具

    公开资料标注支持 Codex、Claude、Cursor、Gemini CLI;先确认自己用的是其中之一,再按对应工具的 Skills 目录放置。

  2. 2

    拿到安装命令

    可以直接复制的命令是 npx skills add https://github.com/anortham/goldfish --skill plan;命令里的占位符换成自己的路径或密钥再执行。

  3. 3

    安装后重启宿主

    命令执行完要完全退出宿主工具再打开,让新装的 Skill 被重新扫描到,否则列表里可能看不到。

  4. 4

    按最小权限试用

    它需要调用外部服务,先在测试目录或测试账号里跑,确认没问题再放到正式项目。

常见问题

6 条

plan 怎么安装?

在终端执行 npx skills add https://github.com/anortham/goldfish --skill plan。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

plan 支持哪些 AI 工具?

公开资料标注支持 Codex、Claude、Cursor、Gemini CLI。其他工具只要能读取本地 Skill 目录,一般也能手动放进去,但要自己确认触发方式是否一致。

plan 需要什么权限?

权限级别是调用外部服务。该 Skill 可能调用第三方服务、云服务或外部模型 API,使用前需要确认账号、额度、数据发送范围和服务条款。

plan 是免费的吗?

公开资料没有标注明确许可证,收费和商用限制都无法确认。正式使用前请到来源仓库核对授权说明。

plan 用的人多吗?

公开数据是累计安装 306 次,近一周安装 13 次,下载 107 次。这些数字来自第三方目录统计,只能当作热度参考,不代表质量或安全性。

plan 安全吗?

第三方扫描结果为通过,但这只覆盖已知的静态检查项,不能替代自己读一遍源码和安装命令。

值得选的地方和要先想清楚的地方

值得选的地方

  • 同时适配 4 种工具,换客户端不用重新找替代品。
  • 第三方扫描已通过,常见的高风险行为没有命中。

先想清楚的地方

  • 许可证没有明确标注,能不能商用、能不能再分发都不确定。
  • 第三方目录整理,版本更新和行为细节以来源仓库为准。

适不适合你

更适合

  • 已经在用 Codex,想直接扩展它能力的人。
  • 需要处理研究检索类任务、又不想每次都手写提示词的人。

不太适合

  • 需要明确授权才能用到公司项目里的场景。
  • 需要厂商级技术支持和 SLA 的生产系统。

研究检索类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
find-skills

寻找技能 · Codex、Claude

31,199,81316,486操作浏览器MIT有提醒
self-improving-agent

自我改进 Agent · OpenClaw

16,464,0023,396接触敏感数据MIT-0通过
soultrace

灵魂追踪 · Codex、Claude

7,588,94412需要联网未明确有提醒
self-improving

自我完善 · OpenClaw

6,829,2841,046需要联网MIT-0通过
microsoft-foundry

微软代工厂 · Codex、Claude

6,614,144764调用外部服务未明确有提醒
plan

计划 · 本页

306—调用外部服务未明确通过

按安装量取同分类前几名;数字来自第三方目录统计,只作热度参考,不代表质量结论。

SKILL.md 原文

Plan — Persistent Strategic Plans

The Core Idea

Plans capture strategic direction that must survive across sessions. When you start multi-session work or make architectural decisions, save a plan immediately — the reasoning is fresh now and won't be available later. When work is done, mark the plan complete so future sessions don't waste effort on finished goals. A stale plan actively misleads; it's worse than having no plan at all.

When to Create a Plan

Plans matter most when context needs to survive session boundaries. Without a plan, the next session starts from scratch.

  • Starting multi-session work — save a plan with goals and approach so the next session knows where to pick up
  • After architectural decisions — capture the direction immediately so future sessions don't re-derive or contradict it
  • After brainstorming/design sessions — the approved design should outlive the conversation that produced it
  • New features spanning multiple sessions — a plan keeps the thread when context compacts or sessions end
  • User describes project direction or roadmap — that strategic context is exactly what plans are for

How to Create a Plan

mcp__goldfish__plan({
  action: "save",
  title: "Auth System Overhaul",
  content: "## Goals\n\n- Migrate from session tokens to JWT...\n\n## Approach\n\n...\n\n## Tasks\n\n- [ ] Task 1\n- [ ] Task 2",
  tags: ["feature", "auth"],
  activate: true
})

Plan Content Structure

Structure your plan content with clear sections:

  1. Goals — what are we trying to achieve?
  2. Approach — how are we going to do it?
  3. Tasks — checklist of deliverables (use - [] / - [x])
  4. Constraints — anything we need to avoid or work around

Key Parameters

  • activate: true — set this so the plan appears at the top of every recall() response. Without activation, future sessions won't see the plan and the strategic direction is effectively lost.
  • tags — categorize for search. Use consistent tags across checkpoints and plans.
  • id — auto-generated from title. Only override if you need a specific slug.

Managing Plan Lifecycle

Check current plans

mcp__goldfish__plan({ action: "list" })

Update a plan (add progress, change scope)

mcp__goldfish__plan({
  action: "update",
  id: "auth-system-overhaul",
  updates: {
    content: "## Goals\n\n...(updated content with progress noted)..."
  }
})

Mark a plan complete

mcp__goldfish__plan({ action: "complete", id: "auth-system-overhaul" })

Archive a superseded plan

mcp__goldfish__plan({
  action: "update",
  id: "old-plan",
  updates: { status: "archived" }
})

The Active Plan

Only ONE plan can be active per workspace. The active plan:

  • Appears at the top of every recall() response
  • Guides all work in the project
  • Should reflect the current strategic direction

If priorities shift, either update the active plan or archive it and create a new one.

Why These Rules Matter

Plans exist so future sessions know the strategic direction. A plan that isn't activated or maintained actively harms future work — it's worse than having no plan at all.

  • Activate plans so they appear in recall(). An inactive plan is invisible to future sessions, which defeats the purpose of creating it.
  • Mark plans complete when done. A stale active plan misleads every future session into thinking work is still in progress and wastes effort re-investigating completed goals.
  • Complete or archive abandoned plans. Orphaned active plans create confusion about what's actually being worked on.
  • Update plans as work progresses. Check off tasks, note scope changes, record decisions — a plan that doesn't reflect reality isn't useful.
  • Plans are forward-looking, checkpoints are backward-looking. Plans capture where you're going. Checkpoints capture where you've been. Use both — they serve different purposes.

平台分布

Codex

37.07%
按下载量换算40

Claude

27.42%
按下载量换算29

Cursor

19.4%
按下载量换算21

Gemini CLI

9.76%
按下载量换算10

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

调用外部服务

该 Skill 可能调用第三方服务、云服务或外部模型 API,使用前需要确认账号、额度、数据发送范围和服务条款。

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。目前只有一个公开来源,建议同时核对源仓库。

继续浏览同类 Skills