Token导航 LogoToken导航

Agent Skill · 研究检索

check-plan(检查计划)

check-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。

check-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,权限为只读;许可证信息还不明确,第三方扫描有提醒项,按安装量在研究检索类的 73,675 个 Skill 里排第 50,054。

总安装 630周安装 26GitHub Stars 公开资料未说明下载 206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:check-plan(检查计划)
来源仓库:https://github.com/cygnusfear/agent-skills
仓库路径:skills/check-plan
安装命令:
npx skills add https://github.com/cygnusfear/agent-skills --skill check-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/cygnusfear/agent-skills --skill check-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。

用于实施计划进度审计和质量验证。可对照计划要求检查已完成工作的完整性和准确性。生成剩余任务的 actionable 清单供后续执行。check-plan 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

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。

来源仓库内路径:skills/check-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/cygnusfear/agent-skills --skill check-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

    它标注为只读,先用只读任务试一次,确认行为符合预期再扩大使用范围。

常见问题

6 条

check-plan 怎么安装?

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

check-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。

check-plan 需要什么权限?

权限级别是只读。该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

check-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。

check-plan 用的人多吗?

公开数据是累计安装 630 次,近一周安装 26 次,下载 206 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。

check-plan 安全吗?

第三方扫描有提醒项,说明存在需要人工确认的行为。安装前请打开源码看清它读了什么、联网到哪里。

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。
  • 标注为只读权限,试用阶段不容易误改本地文件。

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。
  • 第三方扫描存在提醒项或未通过项,需要人工逐条确认。
  • 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。
  • 只想先试试效果、不希望它改动本地文件的人。

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
check-plan

检查计划 · 本页

630—只读未明确有提醒

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

Check Plan Implementation

Instructions

Perform comprehensive audit of implementation progress against a plan, verify quality of completed work, and generate actionable task list for remaining items.

Phase 1: Setup & Discovery

Step 1: Identify the Plan
  • Ask user which plan to check (or identify from context)
  • Read the plan ticket using tk show <ticket-id> (search with tk list --tag plan)
  • Understand all plan items and requirements
Step 2: Get Git Context (if applicable)
# See what files changed
git status

# See detailed changes
git diff

# See commit history on this branch
git log --oneline -20

This helps understand scope of changes made.

Step 3: Identify All Affected Files

Create comprehensive list:

  1. Files mentioned in the plan
  2. Files shown in git status
  3. Files that might be affected (use Glob/Grep)

Create todo list with one item per file to check.

Phase 2: Systematic File-by-File Audit

For EACH file in the todo list:

Step 1: Read the File
  • Use Read tool to examine current state
  • Check memory/context for any previous notes about this file
Step 2: Map to Plan Items

Identify which plan step(s) relate to this file:

  • Which implementation steps mention this file?
  • What changes were supposed to be made?
  • What requirements from the plan apply here?
Step 3: Verify Implementation

Check if planned changes are present:

  • ✅ DONE: Implementation matches plan requirements
  • ⚠️ PARTIAL: Some work done, but incomplete
  • ❌ NOT DONE: No implementation yet
  • 🔍 NEEDS REVIEW: Implementation exists but may not match plan

For each file, assess:

  • Are planned features implemented?
  • Is code quality good?
  • Are types used correctly (no any)?
  • Does it follow AGENTS.md guidelines?
  • Is it complete or partial?
Step 4: Quality Verification

Check implementation quality:

  • Correctness: Does it work as planned?
  • Types: Proper typing, no any, using existing types?
  • Naming: Follows conventions from plan?
  • Architecture: Matches planned design?
  • Completeness: All details from plan step implemented?
Step 5: Record Assessment

Store in memory:

File: path/to/file.ts
Plan Item: Step X - [description]
Status: [DONE|PARTIAL|NOT DONE|NEEDS REVIEW]
Notes: [What's good, what's missing, what needs fixing]
Quality Issues: [Any problems found]
Step 6: Update Todo

Mark file as checked in the todo list.

Phase 3: REMOVAL SPEC Verification

CRITICAL: Verify old code was actually removed.

Step 1: Read REMOVAL SPEC from Plan

Extract all items listed for removal:

  • Code to be removed (with file/line numbers)
  • Files to be deleted
  • Deprecated functions to eliminate
Step 2: Verify Each Removal

For each item in REMOVAL SPEC:

  1. For code to remove: # Check if old code still exists grep -n "old_function_name" path/to/file.ts

- ✅ If not found: Code successfully removed - ❌ If found: Code still exists (NOT DONE)

  1. For files to delete: # Check if file still exists ls path/to/old-file.ts

- ✅ If not found: File successfully deleted - ❌ If found: File still exists (NOT DONE)

  1. For deprecated imports/references: # Search entire codebase grep -r "old_symbol" src/

- ✅ If not found: All references removed - ❌ If found: References still exist (NOT DONE)

Step 3: Record Removal Status
  • List what was supposed to be removed
  • List what actually was removed
  • Flag any items not removed as HIGH PRIORITY tasks

Phase 3.5: Original Issue/Task Coverage Verification (MANDATORY)

CRITICAL: Verify that the implementation covers 100% of the ORIGINAL issue/task requirements, not just the plan steps.

Step 1: Locate Original Issue/Task

Find the source requirement:

  • GitHub issue that triggered this plan: gh issue view <number>
  • Original task description or ticket
  • User request that initiated the work
Step 2: Extract ALL Original Requirements

From the original issue/task, extract:

  • Every functional requirement
  • Every acceptance criterion
  • Every edge case mentioned
  • Every error handling requirement
  • Any implicit requirements
Step 3: Map Requirements to Implementation
#Original RequirementPlan StepImplementation Status
1[from issue]Step X✅/❌/⚠️
2[from issue]Step Y✅/❌/⚠️
3[from issue]N/A (missing from plan!)❌
Step 4: Identify Coverage Gaps

Two types of gaps:

  1. Plan gaps: Requirements from issue that weren't captured in plan
  2. Implementation gaps: Plan steps that weren't fully implemented

Both count toward incomplete coverage.

Step 5: Coverage Assessment
Issue Coverage = (Implemented Original Requirements / Total Original Requirements) × 100%

Anything less than 100% = PLAN NOT COMPLETE

Phase 4: Gap Analysis

Step 1: Identify Scope Creep

Files changed that are NOT in the plan:

  • Why were they changed?
  • Were changes necessary?
  • Should plan be updated to reflect them?
Step 2: Identify Missing Work

Plan items without corresponding implementation:

  • Which steps haven't been started?
  • Which steps are partially complete?
  • What's blocking completion?
Step 3: Identify Issue Coverage Gaps (HIGH PRIORITY)

Requirements from original issue not in implementation:

  • Which issue requirements are missing?
  • Were they missed in planning or implementation?
  • These are HIGHER PRIORITY than plan step completion

Phase 5: Build Progress Report

Step 1: Calculate Completion Percentage
Total Plan Steps: X
Completed Steps: Y
Partial Steps: Z
Not Started: W

Completion: (Y / X) * 100%
Weighted Completion: ((Y + 0.5*Z) / X) * 100%
Step 2: Generate Structured Report

Add a progress note to the plan ticket using tk add-note <ticket-id> "<progress>", or create a separate progress report:

# Plan Progress Report: [Plan Name]
**Date**: [timestamp]
**Plan File**: [path]
**Status**: [In Progress | Ready for Review | Completed]

---

## Summary
- **Overall Completion**: X%
- **Steps Complete**: Y / Total
- **Steps Partial**: Z
- **Steps Not Started**: W
- **Critical Issues**: N

---

## Progress by Plan Step

### ✅ Step 1: [Description]
**Status**: DONE
**Files**: [list]
**Notes**: [Any relevant notes]

### ⚠️ Step 2: [Description]
**Status**: PARTIAL (60% complete)
**Files**: [list]
**Completed**:
- [What's done]
**Remaining**:
- [What's not done]
**Issues**: [Any problems]

### ❌ Step 3: [Description]
**Status**: NOT DONE
**Blocking**: [What's blocking this]

---

## REMOVAL SPEC Status

### ✅ Completed Removals
- `old_function` from `file.ts` - Successfully removed
- `old-file.ts` - Successfully deleted

### ❌ Pending Removals (HIGH PRIORITY)
- `legacy_code` from `file.ts:lines 50-100` - **STILL EXISTS**
- `deprecated-helper.ts` - **FILE STILL EXISTS**

**Critical**: Old code must be removed before plan can be marked complete.

---

## Quality Assessment

### Passed
- ✅ TypeScript types used correctly
- ✅ AGENTS.md naming conventions followed
- ✅ Architecture matches plan

### Issues Found
- ⚠️ `any` type used in `file.ts:42` (should use existing type)
- ⚠️ Missing error handling in step 5 implementation

---

## Files Changed

### Planned Changes (from plan)
- ✅ `path/to/file1.ts` - DONE
- ⚠️ `path/to/file2.ts` - PARTIAL
- ❌ `path/to/file3.ts` - NOT DONE

### Unplanned Changes (scope creep)
- `path/to/unexpected.ts` - Why: [reason]

---

## Remaining Work

### High Priority
1. **Remove old code** (REMOVAL SPEC items)
   - [ ] Remove `legacy_code` from `file.ts`
   - [ ] Delete `deprecated-helper.ts`

2. **Complete Step 3**
   - [ ] Implement [specific requirement]
   - [ ] Add proper types

### Medium Priority
3. **Fix quality issues**
   - [ ] Replace `any` in `file.ts:42`
   - [ ] Add error handling in step 5

### Low Priority
4. **Polish**
   - [ ] [Minor improvements]

---

## Validation Status

### Pre-Validation
- [ ] All plan steps completed
- [ ] All REMOVAL SPEC items removed
- [ ] TypeScript compiles
- [ ] Linting passes
- [ ] No `any` types added

**Ready for Final Validation**: NO (pending items remain)

---

## Next Steps

1. Complete REMOVAL SPEC items (remove old code)
2. Finish Step 3 implementation
3. Fix quality issues
4. Run validation checks
5. Update plan status when 100% complete

Phase 6: Validation Checks

Step 1: Run TypeScript Check
# Run type checking
npm run typecheck
# or
tsc --noEmit

Record results:

  • ✅ Passes: Good to go
  • ❌ Errors: List errors, add to remaining work
Step 2: Run Linting
# Run linter
npm run lint
# or
eslint .

Record results:

  • ✅ Passes: Good to go
  • ⚠️ Warnings: List warnings
  • ❌ Errors: List errors, add to remaining work
Step 3: Run Build (if applicable)
npm run build

Ensure build succeeds.

Phase 7: Generate Task List

Create actionable todo list for remaining work:

# Remaining Tasks for [Plan Name]

## Critical (Must Do)
- [ ] Remove `legacy_code` from `file.ts:50-100` (REMOVAL SPEC)
- [ ] Delete `deprecated-helper.ts` (REMOVAL SPEC)
- [ ] Complete Step 3: [description]

## Important (Should Do)
- [ ] Fix TypeScript error in `file.ts:42`
- [ ] Add error handling in step 5

## Polish (Nice to Have)
- [ ] [Minor improvement]

## Validation
- [ ] TypeScript passes (`npm run typecheck`)
- [ ] Linting passes (`npm run lint`)
- [ ] Build succeeds (`npm run build`)
- [ ] All REMOVAL SPEC items removed

**When all tasks complete**: Close the plan ticket with `tk close <ticket-id>`

Phase 8: Report to User

Provide concise summary:

# Plan Check Complete: [Plan Name]

## Overall Status
**X% Complete** (Y of Z steps done)

## Completed ✅
- Step 1: [description]
- Step 2: [description]

## In Progress ⚠️
- Step 3: [description] (60% done)

## Not Started ❌
- Step 4: [description]

## Critical Issues 🚨
- **REMOVAL SPEC not complete**: Old code still exists
  - `legacy_code` in `file.ts` must be removed
  - `deprecated-helper.ts` must be deleted

## Quality Issues
- `any` type used in `file.ts:42`
- Missing error handling in step 5

## Validation
- ❌ TypeScript: 3 errors
- ✅ Linting: Passed
- Build: Not tested

## Next Steps
1. Remove old code (REMOVAL SPEC)
2. Complete Step 3
3. Fix quality issues
4. Run final validation

**Detailed Report**: Added as note to plan ticket
**Task List**: See remaining work section above

Critical Principles

  • NEVER SKIP FILES - Check every file in the comprehensive list
  • DO NOT EDIT FILES - This is read-only audit, not implementation
  • VERIFY REMOVAL SPEC - Critical that old code is actually removed
  • BE THOROUGH - Think critically about each file's implementation
  • USE MEMORY - Store context as you review files
  • RUN VALIDATION - Always run typecheck and lint
  • BE HONEST - Mark things as incomplete if they are
  • PROVIDE ACTIONS - Don't just identify issues, provide todo list
  • CHECK QUALITY - Implementation exists doesn't mean it's good
  • 100% STANDARD - Plan isn't done until 100% complete and validated

Completion Criteria

A plan can be marked as .done.md ONLY when:

  1. ✅ 100% of ORIGINAL ISSUE/TASK requirements implemented (not just plan steps!)
  2. ✅ All plan steps implemented (100% completion)
  3. ✅ All REMOVAL SPEC items removed (old code gone)
  4. ✅ TypeScript passes (tsc --noEmit succeeds)
  5. ✅ Linting passes (no errors)
  6. ✅ Build succeeds (if applicable)
  7. ✅ No any types added (strict typing maintained)
  8. ✅ AGENTS.md compliance (all guidelines followed)
  9. ✅ Quality verified (implementations match plan specs)

CRITICAL: Criterion #1 is the MOST IMPORTANT. A plan that completes all its steps but doesn't fulfill the original issue requirements is STILL INCOMPLETE.

Anything less = plan ticket stays open

Supporting Tools

  • Grep: Search for old code to verify removal
  • Glob: Find all relevant files
  • Bash: Run git, typecheck, lint, build
  • Read: Examine file contents
  • TodoWrite: Track file review progress
  • Memory/Pinboard: Store context across files

After completing the plan check, follow handbook 15.04 to create tk tickets for all surfaced issues.

平台分布

Codex

33.41%
按下载量换算69

Claude

32.28%
按下载量换算66

Cursor

20.36%
按下载量换算42

Gemini CLI

9.31%
按下载量换算19

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

只读

该 Skill 主要提供规则、说明或参考内容,本身偏只读;真正读写文件、联网或执行命令仍取决于宿主 Agent 的任务。

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。来源安全扫描存在提醒或未通过项,安装前请自行复核。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