Token导航 LogoToken导航

Agent Skill · 研究检索

ark-issues(方舟问题)

用于围绕 GitHub 仓库、Issue、Pull Request、分支、提交和代码协作流程提供辅助能力。它适合让 Agent 查询项目状态、整理变更、辅助创建或检查协作事项,并把仓库中的信息转成可执行的下一步。使用时需要区分只读查询和写入操作;涉及创建 PR、修改 Issue、推送分支或访问私有仓库时,应确认 token 权限、目标仓库范围和用户授权。

ark-issues 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,权限为需要联网;许可证为 MIT,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 47,174。

总安装 745周安装 32GitHub Stars 373下载 261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
MIT

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:ark-issues(方舟问题)
来源仓库:https://github.com/mckinsey/agents-at-scale-ark
仓库路径:skills/ark-issues
安装命令:
npx skills add https://github.com/mckinsey/agents-at-scale-ark --skill ark-issues
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。不同来源提供的安装方式可能略有差异;本站展示可直接复制的安装命令,安装前请核对来源页面。

skills.shnpx skills
npx skills add https://github.com/mckinsey/agents-at-scale-ark --skill ark-issues

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。

ark-issues 管理 Ark 项目的 GitHub Issue 生命周期,支持搜索、创建与关联安全漏洞议题。适合跟踪 CVE 修复进度、链接 PR 到已有问题单,或发起新功能需求讨论时使用。集成 GitHub CLI 命令实现批量操作,如按标签筛选、状态更新与评论追加。

写入操作需验证 PAT token 权限范围,避免误改他人议题或触发速率限制。ark-issues 属于研究检索类 Skill,可作为该场景下的辅助能力补充。

来源仓库内路径:skills/ark-issues

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/mckinsey/agents-at-scale-ark --skill ark-issue…;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

    它需要需要联网,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

ark-issues 怎么安装?

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

ark-issues 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。

ark-issues 需要什么权限?

权限级别是需要联网。该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

ark-issues 是免费的吗?

来源标注的许可证是 MIT。本站不代收任何费用,也不托管安装包;能否商用、要不要署名,以许可证原文为准。

ark-issues 用的人多吗?

公开数据是累计安装 745 次,近一周安装 32 次,下载 261 次,GitHub 373 星。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。

ark-issues 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。
  • 3 个目录同时收录,说明这个 Skill 流传比较广。
  • 许可证明确为 MIT,商用和分发边界好判断。

先想清楚的地方

  • 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。

不太适合

  • 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
ark-issues

方舟问题 · 本页

745373需要联网MIT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

Ark Issues

Manage GitHub issues for the Ark project (mckinsey/agents-at-scale-ark).

When to use this skill

Use this skill when:

  • Searching for existing issues by keyword or CVE number
  • Finding issues related to security vulnerabilities
  • Creating new issues to track bugs or features
  • Viewing issue details and status
  • Listing open or closed issues

Note: This skill is commonly used by the ark-security-patcher agent to:

  1. Search for existing CVE-related issues before starting work
  2. Link PRs to existing issues with "Closes #N" syntax
  3. Create new issues for tracking discovered vulnerabilities

GitHub CLI Commands

Use the gh CLI tool for all issue operations:

Searching Issues

# Search issues by keyword
gh search issues --repo mckinsey/agents-at-scale-ark "CVE"

# Search for specific CVE numbers
gh search issues --repo mckinsey/agents-at-scale-ark "CVE-2025-55183"

# Search with filters
gh search issues --repo mckinsey/agents-at-scale-ark "security" --state open
gh search issues --repo mckinsey/agents-at-scale-ark "vulnerability" --label security

Listing Issues

# List all open issues
gh issue list --repo mckinsey/agents-at-scale-ark

# List issues with filters
gh issue list --repo mckinsey/agents-at-scale-ark --state open
gh issue list --repo mckinsey/agents-at-scale-ark --label bug
gh issue list --repo mckinsey/agents-at-scale-ark --assignee @me

# List with custom fields
gh issue list --repo mckinsey/agents-at-scale-ark --json number,title,state,labels

Viewing Issue Details

# View specific issue
gh issue view 123 --repo mckinsey/agents-at-scale-ark

# View with comments
gh issue view 123 --repo mckinsey/agents-at-scale-ark --comments

# View as JSON for parsing
gh issue view 123 --repo mckinsey/agents-at-scale-ark --json number,title,body,state,labels

Creating Issues

# Create issue interactively
gh issue create --repo mckinsey/agents-at-scale-ark

# Create with title and body
gh issue create --repo mckinsey/agents-at-scale-ark \
  --title "Security: Fix CVE-2025-XXXXX" \
  --body "Description of the vulnerability..."

# Create with labels
gh issue create --repo mckinsey/agents-at-scale-ark \
  --title "Bug: API endpoint fails" \
  --body "Steps to reproduce..." \
  --label bug,priority:high

Updating Issues

# Close an issue
gh issue close 123 --repo mckinsey/agents-at-scale-ark

# Reopen an issue
gh issue reopen 123 --repo mckinsey/agents-at-scale-ark

# Add comment
gh issue comment 123 --repo mckinsey/agents-at-scale-ark \
  --body "Fixed in PR #456"

# Edit issue
gh issue edit 123 --repo mckinsey/agents-at-scale-ark \
  --title "New title" \
  --add-label security

Common Workflows

Workflow 1: Check for Existing CVE Issues

Before creating a new security fix, check if an issue already exists:

# Search for CVE number
gh search issues --repo mckinsey/agents-at-scale-ark "CVE-2025-55183"

# If found, note the issue number
# If not found, you may want to create one

Tip: When creating PRs, reference the issue number using Closes #123 in the PR body to automatically close the issue when the PR merges.

Workflow 2: Find All Security-Related Issues

# Search by keyword
gh search issues --repo mckinsey/agents-at-scale-ark "security OR vulnerability OR CVE"

# Filter by label if security labels exist
gh issue list --repo mckinsey/agents-at-scale-ark --label security

Workflow 3: Create Security Issue for Tracking

gh issue create --repo mckinsey/agents-at-scale-ark \
  --title "fix: CVE-2025-XXXXX in [component]" \
  --body "$(cat <<'EOF'
## Vulnerability Details
- **CVE**: CVE-2025-XXXXX
- **Severity**: High
- **Component**: [package name]

## Description
[What the vulnerability is]

## Impact on Ark
[How it affects Ark]

## Proposed Fix
[Update package to version X.Y.Z]

## References
- CVE: https://cve.circl.lu/cve/CVE-2025-XXXXX
- Advisory: [URL]
EOF
)" \
  --label security

Best Practices

Before Creating Issues

  1. Always search first: Check if a similar issue already exists gh search issues --repo mckinsey/agents-at-scale-ark "keyword"
  2. Be specific: Use clear, descriptive titles

- Good: "fix: CVE-2025-55183 in Next.js affects dashboard" - Bad: "security issue"

  1. Include context: Provide all relevant details in the issue body

When Linking Issues to PRs

  • Use Closes #123 or Fixes #123 in PR descriptions to auto-close issues
  • Reference multiple issues: Closes #123, Closes #456
  • Use issue numbers in commit messages for traceability

Issue Formatting

For security issues, use this template:

## Vulnerability Details
- **CVE**: CVE-YYYY-NNNNN
- **Severity**: [Critical/High/Medium/Low]
- **Component**: [Package/library name]

## Description
[Clear explanation of the vulnerability]

## Impact on Ark
[Which services are affected and how]

## Proposed Fix
[Recommended mitigation approach]

## References
- CVE: [URL]
- Advisory: [URL]

Error Handling

Issue Not Found

gh issue view 999 --repo mckinsey/agents-at-scale-ark
# Error: issue not found

Solution: Verify the issue number is correct

Permission Denied

gh issue create --repo mckinsey/agents-at-scale-ark
# Error: permission denied

Solution: Ensure you're authenticated with gh auth status and have write access to the repo

Rate Limiting

If you hit GitHub API rate limits:

  • Wait a few minutes before retrying
  • Use gh auth status to check your rate limit status
  • Consider batching operations

Integration with Security Workflow

The ark-security-patcher agent uses this skill to:

  1. Search for existing CVE issues before starting work: gh search issues --repo mckinsey/agents-at-scale-ark "CVE-2025-55183"
  2. Link PRs to issues by including in PR body: Closes #33
  3. Track vulnerability fixes by creating issues when CVEs are discovered

Example workflow:

# Agent searches for CVE issue
ISSUE=$(gh search issues --repo mckinsey/agents-at-scale-ark "CVE-2025-55183" --json number --jq '.[0].number')

if [ -n "$ISSUE" ]; then
  echo "Found existing issue #$ISSUE"
  # Include "Closes #$ISSUE" in PR
else
  echo "No existing issue found"
  # Optionally create a new issue
fi

Important Notes

  • Repository: All commands target mckinsey/agents-at-scale-ark
  • Authentication: Requires gh CLI to be authenticated (gh auth login)
  • Permissions: Need read access to search/view, write access to create/update
  • Rate limits: GitHub API has rate limits; be mindful of excessive searches

Common Patterns

Parse JSON Output

# Get issue numbers matching a search
gh search issues --repo mckinsey/agents-at-scale-ark "CVE" \
  --json number,title --jq '.[] | "\(.number): \(.title)"'

# Check if issue exists
EXISTS=$(gh search issues --repo mckinsey/agents-at-scale-ark "CVE-2025-55183" --json number --jq 'length')
if [ "$EXISTS" -gt 0 ]; then
  echo "Issue exists"
fi

Batch Operations

# List all open security issues
gh issue list --repo mckinsey/agents-at-scale-ark \
  --label security --state open \
  --json number,title --jq '.[] | "\(.number): \(.title)"'

# Create multiple issues from a list
for cve in CVE-2025-001 CVE-2025-002; do
  gh issue create --repo mckinsey/agents-at-scale-ark \
    --title "Security: Fix $cve" \
    --body "Track fix for $cve"
done

平台分布

Claude Code

29%
按下载量换算76

windsurf

23.48%
按下载量换算61

OpenCode

16.84%
按下载量换算44

Codex

11.48%
按下载量换算30

Antigravity

6.42%
按下载量换算17

Gemini CLI

3.14%
按下载量换算8

安全审计

Gen Agent Trust Hub

通过

Socket

通过

Snyk

可疑

权限和风险

需要联网

该 Skill 可能需要联网访问来源站点、仓库或外部 API;具体网络访问范围需要结合源码和 README 复核。

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。

继续浏览同类 Skills