Token导航 LogoToken导航

Agent Skill · 研究检索

ln-911-github-triager(LN 911 GitHub triager 搜索)

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。

ln-911-github-triager 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 73,674 个 Skill 里排第 32,401。

总安装 2,521周安装 104GitHub Stars 441下载 824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:ln-911-github-triager(LN 911 GitHub triager 搜索)
来源仓库:https://github.com/levnikolaevich/claude-code-skills
仓库路径:skills/ln-911-github-triager
安装命令:
npx skills add https://github.com/levnikolaevich/claude-code-skills --skill ln-911-github-triager
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/levnikolaevich/claude-code-skills --skill ln-911-github-triager

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 GitHub 仓库 Issue、PR 和讨论的自动化管理与优先级排序。支持按时间、类型和标签过滤未处理事项。提供红旗指标和参与度分析辅助决策。使用时需确认 gh CLI 权限和目标仓库访问范围。ln-911-github-triager 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/ln-911-github-triager

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/levnikolaevich/claude-code-skills --skill ln-9…;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

ln-911-github-triager 怎么安装?

在终端执行 npx skills add https://github.com/levnikolaevich/claude-code-skills --skill ln-911-github-…。装完完全退出宿主工具再打开,让它重新扫描本地 Skill。

ln-911-github-triager 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。

ln-911-github-triager 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。

ln-911-github-triager 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。

ln-911-github-triager 用的人多吗?

公开数据是累计安装 2,521 次,近一周安装 104 次,下载 824 次,GitHub 441 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。

ln-911-github-triager 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。

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。

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
ln-911-github-triager

LN 911 GitHub triager 搜索 · 本页

2,521441需要联网未明确有提醒

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

Paths: File paths (shared/, references/, ../ln-*) are relative to skills repo root. If not found at CWD, locate this SKILL.md directory and go up one level for repo root. If shared/ is missing, fetch files via WebFetch from https://raw.githubusercontent.com/levnikolaevich/claude-code-skills/master/skills/{path}.

ln-911-github-triager

Type: L3 Worker (standalone) Category: 9XX Community Engagement Invocation: Standalone or as part of a broader community triage workflow

Scans all open issues, PRs, and discussions in the current repository. Categorizes by urgency. Produces an actionable triage report.


Arguments

$ARGUMENTS controls scope and staleness threshold:

TokenEffectDefault
issuesScan issues onlyAll three
prsScan PRs onlyAll three
discussionsScan discussions onlyAll three
Number (e.g., 14)Staleness threshold in days7
issues 30Combine: scope + threshold--
summaryCompact output for coordinator: skip Phase 4 preview, return red flags + priority counts + metrics onlyFull interactive report

Parse $ARGUMENTS: extract scope tokens (issues, prs, discussions), numeric threshold, and summary flag. If no scope tokens, scan all three. If no number, default to 7 days.


Phase 0: GitHub Discovery

MANDATORY READ: Load shared/references/community_github_discovery.md

Execute the discovery protocol. Extract:

  • {owner}/{repo} for --repo flag and URLs
  • maintainer login (authenticated user) for maintainer detection
  • Verify gh is authenticated

Load strategy: check docs/community_engagement_strategy.md in target project, fallback to shared/references/community_strategy_template.md. Extract Section 5 (Engagement Metrics) for targets and red flags.


Phase 1: Fetch Data

Run these gh commands based on scope. Execute all applicable fetches in parallel.

1a. Issues

gh issue list --repo {owner}/{repo} --state open --limit 100 --json number,title,labels,createdAt,updatedAt,author,comments,assignees

1b. Pull Requests

gh pr list --repo {owner}/{repo} --state open --limit 100 --json number,title,labels,createdAt,updatedAt,author,reviewDecision,reviewRequests,isDraft,comments,assignees

1c. Discussions (GraphQL)

gh api graphql -f query='
  query($owner: String!, $name: String!) {
    repository(owner: $owner, name: $name) {
      discussions(first: 100, states: OPEN) {
        nodes {
          number
          title
          category { name }
          createdAt
          updatedAt
          author { login }
          answerChosenAt
          labels(first: 5) { nodes { name } }
          comments(first: 20) {
            totalCount
            nodes {
              author { login }
              createdAt
            }
          }
        }
      }
    }
  }
' -f owner="{owner}" -f name="{repo}"

Gate: If all sources return zero items, report "No open items. Triage complete." and stop.


Phase 2: Classify Each Item

For every fetched item, evaluate these conditions. An item can match multiple -- use highest priority.

Priority Matrix

PriorityConditionApplies to
P0 CriticalLabel contains bug AND no assigneeIssues
P0 CriticalLabel contains breakingIssues, PRs
P1 UrgentAuthor!= maintainer AND zero comments from maintainerIssues, Discussions
P1 UrgentQ&A discussion with answerChosenAt is null AND author!= maintainerDiscussions
P1 UrgentreviewDecision is null or REVIEW_REQUIREDPRs
P2 StaleNo activity for > threshold days (updatedAt older than threshold)All
P3 MonitorPR marked as draft (isDraft is true)PRs
P3 MonitorAuthor == maintainer AND 0 community repliesDiscussions
P4 InfoItems not matching any above conditionAll

Maintainer Detection

The maintainer login comes from Phase 0 discovery (viewer.login). An item "has maintainer reply" if any comment's author.login equals the maintainer login.

For issues: check comments array authors. For discussions: check comments.nodes[].author.login.

Staleness Calculation

Days since last activity: (today - updatedAt) in days. Compare against threshold (default 7, or from $ARGUMENTS).

Label Urgency Signals

Labels that elevate priority: bug, breaking, security, urgent.

Labels that signal engagement opportunity (not urgency): help wanted, good first issue.


Phase 3: Engagement Metrics

Calculate aggregate metrics per strategy Section 5:

MetricCalculationTargetRed Flag
Time to First ResponseAvg(first maintainer comment - item created) for items WITH reply<24h>72h
Unanswered discussionsCount where author!= maintainer AND no maintainer comment AND age > threshold0Any >7 days
Community-to-maintainer ratioNon-maintainer comments / maintainer comments across all items>0.30
New discussions this monthDiscussions created in last 30 days>20
Zero-engagement itemsItems with 0 comments total----

Phase 4: Preview Table

If $ARGUMENTS contains summary → skip this phase entirely. Output compact summary (red flags, priority counts per category, health metrics table) and stop. Do NOT compose the full Phase 5 report.

Otherwise (standalone mode): present a compact summary table of ALL classified items for user review. Wait for user approval before proceeding to Phase 5.

### Triage Preview -- {total count} items found

| # | Type | Title | Priority | Age | Key Signal |
|---|------|-------|----------|-----|------------|
| {number} | {Issue/PR/Discussion} | {title (truncated to 50 chars)} | {P0/P1/P2/P3/P4} | {days}d | {short reason} |

Sort by priority (P0 first), then by age (oldest first) within each priority group.

After presenting, ask the user:

  • Confirm priorities look correct
  • Whether to exclude any items from the full report
  • Whether to adjust the staleness threshold

Proceed to Phase 5 only after user confirms.


Phase 5: Compose Report

Present the full triage report in this format:

## GitHub Triage Report -- {YYYY-MM-DD}

**Scope:** {issues/PRs/discussions or "all"} | **Threshold:** {N} days | **Open items:** {total count}

Red Flag Alerts

If any red flags from Phase 3 are triggered, add BEFORE Health Metrics:

> [!WARNING]
> **Red flags detected:**
> - {description, e.g., "2 unanswered discussions older than 7 days"}

Health Metrics Table

| Metric | Value | Target | Status |
|--------|-------|--------|--------|
| Unanswered (>{N}d) | {count} | 0 | {OK / RED FLAG} |
| Avg first response | {hours}h | <24h | {OK / RED FLAG} |
| Community:maintainer ratio | {ratio} | >0.3 | {OK / RED FLAG} |
| New discussions (30d) | {count} | >2 | {OK / RED FLAG} |
| Zero-engagement items | {count} | -- | {count} |

Priority Tables

For each priority P0..P3, show header with count. If count is 0, show header (0) and skip table.

P0 Critical:

#TypeTitleLabelsAgeAction Needed

P1 Urgent:

#TypeTitleAuthorAgeWhy

P2 Stale:

#TypeTitleLast ActivityDays Stale

P3 Monitor:

#TypeTitleNote

P4 Info: Do NOT list individually -- only include in total count.

Recommended Actions

### Recommended Actions

1. **Respond:** {items needing maintainer reply, with GitHub URLs}
2. **Review:** {PRs needing review, with GitHub URLs}
3. **Close or bump:** {stale items -- suggest close if resolved, bump if relevant}
4. **Engage:** {items that could benefit from community call-to-action}

Formatting Rules

  • Sort items within each priority group by age (oldest first)
  • URL format: https://github.com/{owner}/{repo}/{issues|discussions|pull}/{number}
  • Dates in YYYY-MM-DD, ages in days
  • All items in Recommended Actions must have direct GitHub URLs

Rules

  • Read-only: this skill fetches and reports -- never create, modify, or close any GitHub item
  • No publishing: do not create issues, comments, labels, or any other GitHub state change
  • Maintainer login: from Phase 0 discovery (authenticated user)
  • Default threshold: 7 days (per strategy Section 5)
  • Rate limits: if GraphQL rate limit hit, report partial results with a warning

Definition of Done

  • GitHub data fetched for all scoped item types (issues/PRs/discussions)
  • Every item classified by priority (P0-P4)
  • Engagement metrics calculated per strategy targets
  • Report composed (full interactive report OR compact summary for coordinator)

Version: 1.0.0 Last Updated: 2026-03-13

平台分布

Codex

34.43%
按下载量换算284

Claude

29.05%
按下载量换算239

Cursor

17.99%
按下载量换算148

Gemini CLI

9.27%
按下载量换算76

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。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