Token导航 LogoToken导航TokenDH.com
研究检索external-servicegithub未标认证来源可访问许可证需确认审计提醒

catchupcatchup 搜索

Agent Skill

catchup 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

4,480

周安装

183

GitHub Stars

公开资料未说明

下载量

1,449
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:catchup(catchup 搜索)
来源仓库:https://github.com/camacho/ai-skills
仓库路径:skills/catchup
安装命令:
npx skills add https://github.com/camacho/ai-skills --skill catchup
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/camacho/ai-skills --skill catchup

简介

catchup 专注于从代码库和历史记录中提取上下文信息,帮助重建项目进展与任务状态,适合深度代码理解场景。

  • 主要用于读取 git 日志、工作树状态及内存文件,辅助判断当前开发阶段、阻塞项与待办计划。
  • 全程只读操作,不会修改文件或创建分支,可在任意 git 项目中安全使用以获取项目快照。
  • 安装后需确保当前路径下有有效的 git 仓库,并能访问 .branch-context.md 或 ai-workspace 相关文件。
  • 适用于需要快速掌握项目最新状态、识别未完成任务或准备交接的场景,尤其适合代码评审与知识传递。

SKILL.md

Catchup — Context Reconstruction

This skill is strictly read-only. Do NOT edit files, create branches, or run write commands.

1. Gather data (run all independent steps in parallel)

  • git log --format="%h %s (%ar)" -20, git status, and git worktree list
  • Read .branch-context.md (if exists — check CWD and worktree root). Parse YAML frontmatter: workflow_step, validate_round/validate_max, review_round/review_max, plan_review_chunk/plan_review_total/plan_chunks_approved
  • Read ai-workspace/MEMORY.md — note anything in-progress or blocking
  • Read HANDOFF.md (if exists)
  • List ai-workspace/plans/ (sort by mtime, most recent first) and read the first file that does NOT end in .done.md or .executed.md
  • Search Basic Memory via MCP for recent entries matching the current branch name. If MCP unavailable, skip silently.
  • GitHub — check availability first: command -v gh >/dev/null 2>&1 && gh auth status >/dev/null 2>&1 If authenticated, run in parallel: gh pr list --state open gh issue list --state open --limit 15 If gh is unavailable or unauthenticated, note "GitHub CLI unavailable" and skip. Do NOT attempt to install or authenticate.

2. Synthesize

Present findings in this format. Omit sections where the data source was not found (e.g., no HANDOFF.md, gh unavailable). Always show Loop state and Blockers even when empty — absence is informative:

Branch:         <name> — <what we're building (from plan or branch-context)>
Worktree:       <path> (or "primary")
Last commit:    <hash> <message>
Uncommitted:    <staged/unstaged summary, or "clean">
Workflow step:  <N> — <step name> (from frontmatter, or inferred)
Loop state:     validate <round>/<max>, review <round>/<max>
Active plan:    <file> — <current phase> — <next action>
Blockers:       <list>
Open PRs:       <count — titles>
Open issues:    <count — top 5 titles>
Handoff:        <detail>

After the summary, surface blockers and TODO items prominently — remaining work, failing CI, stale branches, items from memory flagged as blocking. These go above the fold, not as footnotes.

3. Suggest next actions

Based on gathered state, suggest 2-3 concrete next actions. Examples:

  • "Resume building — plan X is at step 5 (Build)"
  • "Run /validate — implementation looks complete"
  • "Address blocker: [description]"
  • "Review open PR #N — awaiting action"
  • "Pick up issue #N — in-progress"

End with: "What would you like to do?"

Failure modes

  • No .branch-context.md: infer workflow step from git state and plan file status. Note: "No branch context — step inferred."
  • No active plan: note "No active plan." Do not search .done.md files.
  • On main branch: skip branch-context and plan lookup. Show git state, PRs, issues, and memory only.
  • Basic Memory MCP unavailable: skip silently.
  • gh unavailable: show "GitHub CLI unavailable — PR/issue status unknown."

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

33.72%
按下载量换算489

Claude

33.71%
按下载量换算488

Cursor

18.14%
按下载量换算263

Gemini CLI

10.21%
按下载量换算148

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

external-service

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。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