Token导航 LogoToken导航TokenDH.com
待分类敏感数据github未标认证来源可访问许可证需确认审计通过

plan-status计划状态

Agent Skill

plan-status 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。

总安装

253

周安装

12

GitHub Stars

公开资料未说明

下载量

97
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:plan-status(计划状态)
来源仓库:https://github.com/anortham/goldfish
仓库路径:skills/plan-status
安装命令:
npx skills add https://github.com/anortham/goldfish --skill plan-status
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-status

简介

plan-status 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。

  • 它可辅助查询项目进展、分析代码差异、跟踪 Issue 状态或协助生成协作摘要。
  • 通过 npx skills add 命令从指定仓库安装,具体用法请参考原始 README。
  • 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。
  • 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

Plan Status — Progress Assessment

What This Does

Pulls plans from multiple sources and recent checkpoints, then assesses how actual work aligns with planned goals. Reports what's done, what's next, and whether the project is drifting from the plan.

How to Assess Plan Status

Step 1: Gather plans from ALL sources

Source 1: Goldfish plans

mcp__goldfish__recall({ limit: 0 })

This returns the active Goldfish plan for the current workspace.

If you need a specific plan by ID (not just the active one):

mcp__goldfish__plan({ action: "get", id: "plan-id" })

If no active plan exists, note it but continue — there may still be plans in docs/plans/.

Source 2: Project plan docs

Scan docs/plans/*.md in the current project directory. Read each file and check:

  • The **Status:** field in the header (typically "Approved", "Complete", "In Progress")
  • The date in the filename (format: YYYY-MM-DD-<name>.md)
  • Skip files with Status "Complete" unless they were completed very recently (last 3 days)

Step 2: Recall recent checkpoints

mcp__goldfish__recall({ days: 7, limit: 20, full: true })

Use full: true to get file lists and git metadata — this helps match checkpoints to plan items. Use a wider window (7 days) to capture the full arc of plan execution.

If the plans are older, extend the range:

mcp__goldfish__recall({ days: 14, limit: 30, full: true })

Step 3: Cross-reference and assess

Map each checkpoint to a plan goal. Identify:

  • Which plan items have checkpoint evidence of completion
  • Which plan items have WIP checkpoints but aren't done
  • Which plan items have zero checkpoint activity
  • Which checkpoints don't map to any plan item (scope drift)

For docs/plans files: Do NOT blindly trust the Status header. Verify against checkpoints:

  • Status "Approved" + all tasks have checkpoint evidence → effectively complete
  • Status "Approved" + partial checkpoint coverage → in progress
  • Status "Approved" + no checkpoints → not started

For Goldfish plans: Use the plan's status field directly.

Report Format

Start with a header that identifies the plan and assessment date, then structure as four sections followed by an overall health assessment.

Header

## Plan Status — "Auth System Overhaul" — Feb 14, 2026
Source: Goldfish active plan
3/5 items complete (60%)

When reporting on docs/plans:

## Plan Status — "v5.1 Skills Refresh" — Feb 14, 2026
Source: docs/plans/2026-02-16-v5.1-implementation.md
2/8 tasks complete (25%)

Always include the source location and completion fraction for instant comprehension.

Completed

Plan items with clear checkpoint evidence of completion. Include the approximate date completed and any notable details.

  • JWT refresh token rotation — completed Feb 12, shipped with full test coverage
  • Rate limiter Redis support — completed Feb 13, integration tests passing

In Progress

Plan items with recent checkpoints showing active work but not yet complete.

  • Session management API — 3 checkpoints this week, endpoints implemented but missing error handling
  • Admin dashboard auth — started Feb 13, basic scaffold in place

Not Started

Plan items with no checkpoint activity at all. Flag these — they might be blocked, deprioritized, or forgotten.

  • API documentation — no activity found
  • Load testing — no activity found (may be blocked by staging environment)

Drift Assessment

Work captured in checkpoints that doesn't map to any plan item. This isn't automatically bad — emergent work happens — but it should be called out.

Unplanned work detected:

  • Bug fix: file corruption on concurrent writes (2 checkpoints, Feb 11)
  • Dependency upgrade: fuse.js v7 migration (1 checkpoint, Feb 12)

These consumed ~1 day of effort outside the plan scope.

Overall: Minor Drift

Render the health assessment as a ### Overall: header with the verdict, followed by a direct, honest summary.

Health levels:

  • On track — Most plan items progressing, minimal drift, no blockers
  • Minor drift — Some unplanned work but plan items still advancing
  • Significant drift — More unplanned work than planned work, or key items stalled
  • Stalled — Little to no progress on plan items, or major blockers

Be direct. "The plan called for 5 deliverables this sprint. Two are done, one is in progress, and two haven't been touched. You're behind, and the unplanned auth bug ate a day." That's more useful than "progress is being made."

Multiple Plans

When both a Goldfish active plan AND docs/plans files exist, report on each separately with clear source attribution. Present the Goldfish plan first (it's the "active" strategic direction), then docs/plans (implementation plans).

If plans overlap or conflict, flag it: "The Goldfish active plan focuses on auth, but docs/plans shows active work on a performance benchmark suite. These may be competing priorities."

Critical Rules

  • Do NOT sugarcoat. If the plan is behind, say so. The user needs accurate information, not comfort.
  • Do NOT fabricate progress. Only claim completion for items with actual checkpoint evidence.
  • DO flag scope drift. Unplanned work is a leading indicator of timeline slip.
  • DO suggest plan updates. If a plan is clearly outdated, recommend updating it — via mcp__goldfish__plan({action: "update"}) for Goldfish plans, or by editing the file directly for docs/plans.
  • Match checkpoints to plan items carefully. A checkpoint about "auth" doesn't automatically satisfy a plan item about "auth" — read the descriptions and match on actual content.
  • Include time estimates when possible. "3 checkpoints over 2 days" gives the user a sense of effort invested.
  • Check BOTH plan sources. Missing a source means an incomplete assessment.
  • Attribute sources clearly. The user should always know whether a plan came from .memories/plans/ or docs/plans/.

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

35.74%
按下载量换算35

Claude

26.88%
按下载量换算26

Cursor

18.06%
按下载量换算18

Gemini CLI

9.46%
按下载量换算9

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

通过

权限和风险

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。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