Token导航 LogoToken导航

Agent Skill · 运维和基础设施

standup(standup 搜索)

standup 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息,适合在 Codex、Claude、Cursor、Gemini CLI 中需要围绕仓库状态、代码变更或协作事项进行整理时使用。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。

standup 是运维和基础设施类 Agent Skill,可用宿主为 Codex、Claude、Cursor、Gemini CLI,通过GitHub 仓库获取,权限为接触敏感数据;许可证信息还不明确,第三方扫描已通过,按安装量在运维和基础设施类的 2,449 个 Skill 里排第 2,177。

总安装 256周安装 11GitHub Stars 公开资料未说明下载 90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:standup(standup 搜索)
来源仓库:https://github.com/anortham/goldfish
仓库路径:skills/standup
安装命令:
npx skills add https://github.com/anortham/goldfish --skill standup
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 standup

这个 Skill 是做什么的

运维和基础设施类 Skill 处理部署、监控、日志和云资源,多数需要较高的执行权限。

standup 用于处理 GitHub 仓库、Issue、Pull Request 和代码协作信息。适合在 Codex、Claude、Cursor、Gemini CLI 中围绕仓库状态、代码变更或协作事项进行整理。生成跨项目状态报告,汇总已完成工作、下一步计划与阻塞问题。

安装命令:npx skills add https://github.com/anortham/goldfish --skill standup。需确认权限范围、维护状态及是否触发联网或命令执行。

来源仓库内路径:skills/standup

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 standup;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

    它需要接触敏感数据,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

standup 怎么安装?

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

standup 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。

standup 需要什么权限?

权限级别是接触敏感数据。该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

standup 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。

standup 用的人多吗?

公开数据是累计安装 256 次,近一周安装 11 次,下载 90 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。

standup 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

  • 需要接触敏感数据,装到常用环境前要先评估影响面。
  • 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。
  • 需要处理运维和基础设施类任务、又不想每次都手写提示词的人。

不太适合

  • 环境里有大量密钥、凭证,不方便给第三方 Skill 读取的人。
  • 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。

运维和基础设施类 Skill 对比

6 个同类
Skill总安装GitHub 星权限许可证扫描
brainstorming

头脑风暴 · Codex、Claude

3,148,776173,131操作浏览器MIT通过
copywriting

文案写作 · Codex、Claude

1,996,72425,583需要联网MIT通过
openclaw-secure-linux-cloud

OpenClaw secure linux cloud 命令行 · Codex、Claude

1,853,08840接触敏感数据未明确未通过
xget

xget 命令行 · Codex、Claude

1,799,28040执行命令未明确未通过
requesting-code-review

代码审查请求 · Codex、Claude

1,590,144173,103权限未标注MIT通过
standup

standup 搜索 · 本页

256—接触敏感数据未明确通过

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

Standup — Cross-Project Status Report

What This Does

Generates a concise standup report by recalling checkpoints across all workspaces and reviewing active plans from multiple sources. Covers what happened, what's next, and what's stuck.

How to Generate a Standup

Step 1: Recall cross-project activity

mcp__goldfish__recall({ workspace: "all", days: 1 })

For a Monday standup covering the weekend:

mcp__goldfish__recall({ workspace: "all", days: 3 })

For a custom range:

mcp__goldfish__recall({ workspace: "all", from: "2026-02-10", to: "2026-02-14" })

Step 2: Gather plans from ALL sources

Plans may exist in two locations per project. Check BOTH:

Source 1: Goldfish plans — already included in recall output if an active plan exists.

Source 2: Project plan docs — scan docs/plans/*.md in each project that appeared in the recall results. Read files modified in the last 14 days (skip older ones — they're likely completed). Look at the **Status:** field in the header of each file.

Step 3: Assess plan status from evidence

Do NOT blindly trust the Status field in plan docs. Cross-reference against checkpoints:

  • Status says "Approved" + checkpoints show all tasks done → effectively complete, note as done
  • Status says "Approved" + checkpoints show partial progress → in progress, report remaining items
  • Status says "Approved" + no checkpoint activity → upcoming work, report as planned
  • Status says "Complete" → trust it, skip or mention briefly as done

For Goldfish plans (from recall), use the plan's status field directly — it's managed by the agent via the plan tool.

Step 4: Synthesize the report

Report Format

Structure the standup with these rules:

  • One bullet per accomplishment. Each on its own line for easy scanning.
  • Lead with impact, not activity. "Shipped auth refresh tokens" not "Worked on auth."
  • Use past tense for done items. "Fixed," "Implemented," "Shipped."
  • Blockquote (>) for forward-looking items. Visually separates past from future.
  • Abbreviated month names in headers. Feb 14 not February 14.
  • Date range in header when covering multiple days (Feb 12–14, 2026).
  • Always state blockers explicitly. If none, say "Nothing currently blocked."

Multi-Project Format

When checkpoints span multiple projects, group by project with bullets for accomplishments, blockquotes for next/blocked, and plan references:

## Standup — Feb 14, 2026

### goldfish
- Implemented 4 skill files with behavioral language patterns
- Converted tool handlers from JSON to markdown output
> Next: Test skills with live agent sessions
> Plan: v5.1 skills refresh — 2/4 tasks complete (docs/plans/2026-02-16-v5.1-implementation.md)

### api-gateway
- Fixed rate limiter race condition in Redis cluster mode
- Added integration tests for multi-node scenarios
> Next: Deploy to staging, run load tests
> Blocked: Waiting on DevOps for staging Redis cluster

Single-Project Format

When all checkpoints are from a single project, drop the project grouping and use section headers:

## Standup — Feb 14, 2026

### Done
- Implemented 4 skill files with behavioral language patterns
- Converted tool handlers from JSON to markdown output

### Up Next
- Test skills with live agent sessions
- v5.1 skills refresh: workspace env var implementation (2/4 tasks complete)

### Blocked
- Nothing currently blocked

Synthesis Rules

  • Be concise. A standup is 2 minutes, not 20. One line per accomplishment.
  • Group by theme. Five checkpoints about "auth" become one bullet: "Implemented auth token refresh with rotation."
  • Highlight blockers prominently. These are the most actionable items in a standup.
  • Skip noise. Minor refactors, formatting changes, and config tweaks don't need individual mentions unless they were the main work.
  • Use past tense for accomplishments. "Shipped," "Fixed," "Implemented" — not "Working on" for done items.
  • Surface decisions. If a checkpoint captured an architectural decision, mention it briefly — the team may need to know.
  • Include plan progress. When active plans exist, include a brief progress summary (e.g., "3/5 tasks complete") in the Up Next section.

Handling Edge Cases

No checkpoints found

Report honestly: "No activity recorded in the requested period." Don't fabricate.

Single project only

Use the single-project format above (Done / Up Next / Blocked sections).

Too many checkpoints (20+)

Be more aggressive about grouping. Summarize by theme rather than listing individual items. A standup with 15 bullet points defeats the purpose.

No plans found

That's fine — just skip the plan progress lines. Not every project has active plans.

Plans in docs/plans/ but no Goldfish plan

Include them in the forward-looking section. Plans don't need to be in Goldfish to be useful for standup.

Critical Rules

  • Do NOT ask the user what to include. Recall gives you everything. Synthesize it yourself.
  • Do NOT fabricate activity. Only report what checkpoints actually show.
  • Keep it standup-length. If your report is more than a screenful, you're being too verbose.
  • Include dates when covering multi-day ranges so the reader knows the timeline.
  • Check BOTH plan sources. .memories/plans/ (via recall) AND docs/plans/ (via file reading). Missing one source means an incomplete forward-looking view.
  • Infer plan status from evidence. Don't trust stale Status headers — verify against checkpoint activity.

平台分布

Codex

33.91%
按下载量换算31

Claude

31.13%
按下载量换算28

Cursor

19.14%
按下载量换算17

Gemini CLI

8.74%
按下载量换算8

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

接触敏感数据

该 Skill 可能接触密钥、Token、环境变量或敏感配置,应进入高风险复核队列,默认不自动发布。

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