bug-triage

作者: apify

为Apify MCP服务器分诊开放的GitHub bug问题。获取问题,分析根本原因,起草开发者对开发者的回复,并在用户…后发布。

npx skills add https://github.com/apify/apify-mcp-server --skill bug-triage

Bug Triage

Triage open bug issues on apify/apify-mcp-server. Analyze, draft responses, get approval, post.

Permissions

  • Read-only by default: listing issues (gh issue list), viewing details and comments (gh issue view) — do these automatically, no need to ask.
  • Write requires approval: commenting (gh issue comment), closing (gh issue close), editing (gh issue edit) — always present the draft and wait for explicit user approval before executing.

Step 0: Parse arguments

$ARGUMENTS controls scope:

InputBehavior
623 641 639Triage specific issue numbers
--label bugTriage all open issues with label (default: bug)
--allTriage all open bug-labeled issues
(empty)Same as --all

Step 1: Fetch issues

Use the arguments parsed in Step 0 to determine what to fetch.

If specific issue numbers were provided, skip the list queries and fetch each issue directly:

gh issue view <number> --repo apify/apify-mcp-server --comments

Otherwise (label mode, --all, or empty), fetch open bug reports. Use the label from Step 0 (default: bug). Bug reports come in two forms — find both:

  1. By label: issues with the parsed label
  2. By title prefix: issues with [Bug] in the title (the bug report template adds this, but the label isn't always applied)
# By label (use the label from Step 0, default: bug)
gh issue list --repo apify/apify-mcp-server --label <label> --state open --json number,title,labels,createdAt,body --limit 30

# By title prefix (catches unlabeled bug reports)
gh issue list --repo apify/apify-mcp-server --state open --search "[Bug] in:title" --json number,title,labels,createdAt,body --limit 30

Merge both lists and deduplicate by issue number.

For each issue, fetch full details including all comments (comments often contain the real context):

gh issue view <number> --repo apify/apify-mcp-server --comments

Always read comments before drafting a response. Previous team members may have already replied, asked for details, or provided workarounds. Don't duplicate existing responses.

Step 2: Categorize each issue

Assign one of these categories:

CategoryCriteriaTypical action
Known fixRoot cause is clear, documented fix existsDraft response with fix, optionally close
Not our bugServer logs show it works; issue is in the MCP client (Claude Desktop, Cowork, etc.)Explain findings, suggest workaround
DuplicateSame root cause as another open issueComment with link, close as duplicate
Needs infoNo logs, no config, vague descriptionAsk for specifics
StaleWe asked for info, no reply for 2+ weeksClose with helpful pointer
Actionable bugReal server-side bug we need to fixAcknowledge, investigate further

Common root causes (from past triage)

These patterns recur. Check for them first:

  1. SSE endpoint removed (April 1, 2026): Users on /sse URLs get connection failures. Fix: switch to https://mcp.apify.com (streamable HTTP).

  2. "Tools connected but Claude ignores them": MCP server logs show successful handshake and tools/list, but Claude doesn't use the tools in conversation. This is a Claude Desktop/Cowork behavior, not an Apify server bug. Users need to explicitly ask Claude to use Apify tools.

  3. "Unable to connect to extension server": Claude Desktop extension install issues. Usually caused by corrupted npx cache, Claude silently downgrading the extension, or Node.js not on the GUI app's PATH. Fix: use the remote server https://mcp.apify.com as a custom connector.

  4. npx cache corruption: Stale cache prevents stdio server from starting. Fix: rm -rf ~/.npm/_npx (macOS/Linux) or rmdir /s /q %LOCALAPPDATA%\npm-cache\_npx (Windows).

  5. Claude Desktop connector downgrade: Claude Desktop silently downgrades connectors to older versions. Fix: remove and re-add the connector.

Step 3: Draft response

Style rules:

  • Developer-to-developer, plain language
  • Concise — 3-8 lines typical, no fluff
  • Acknowledge the issue and apologize briefly if the user is stuck
  • Reference documentation when a fix or troubleshooting guide exists
  • If the server works correctly (logs prove it), say so clearly
  • If closing, explain why and point to where to go next

Key documentation links:

  • Setup guide: https://docs.apify.com/platform/integrations/mcp
  • Claude Desktop troubleshooting: https://docs.apify.com/platform/integrations/claude-desktop#troubleshooting
  • Configurator: https://mcp.apify.com

Use Apify MCP tools (search-apify-docs, fetch-apify-docs) to find the right doc link if the issue touches a topic not covered above.

The default recommendation for connection issues is the remote server:

Add a custom connector in Claude Desktop with URL https://mcp.apify.com and follow the OAuth flow.

Step 4: Present and get approval

Present each issue to the user with:

  1. Issue number + title + link (https://github.com/apify/apify-mcp-server/issues/<number>)
  2. Category (from Step 2)
  3. Summary — one line on what's going on
  4. Proposed response — the draft comment in a blockquote
  5. Proposed action — comment only / comment + close / close as duplicate of #X / skip

MANDATORY: Wait for explicit user approval before posting anything. Never post, close, or modify an issue without the user saying yes.

Step 5: Post

After approval:

# Comment only
gh issue comment <number> --repo apify/apify-mcp-server --body "<response>"

# Comment + close
gh issue comment <number> --repo apify/apify-mcp-server --body "<response>" && \
gh issue close <number> --repo apify/apify-mcp-server

# Close as duplicate
gh issue comment <number> --repo apify/apify-mcp-server --body "Closing as a duplicate of #<other> — same root cause. We'll track and update there." && \
gh issue close <number> --repo apify/apify-mcp-server

Then move to the next issue. After the last one, print a summary table:

IssueCategoryAction taken
#623Known fixCommented, SSE migration
#639DuplicateClosed as dup of #641
.........

来自 apify 的更多技能

apify-influencer-brand-collabs
apify
通过串联Apify Actors发现Instagram品牌与创作者的合作关系。当用户询问某品牌与谁合作、某创作者曾与哪些品牌进行付费合作时使用…
apify-actor-development
apify
创建、调试和部署用于网页抓取、自动化及数据处理的无服务器云程序。支持JavaScript、TypeScript和Python模板,集成Crawlee、Playwright和Cheerio库,用于HTTP和基于浏览器的爬取。包含通过apify run进行的本地测试(使用隔离存储)、输入/输出的模式验证,以及通过apify push部署到Apify平台。需要Apify CLI认证,并在.actor/actor.json中强制包含generatedBy元数据以用于AI...
apify-actorization
apify
将现有项目转换为无服务器Apify Actors,支持语言特定的SDK集成。支持JavaScript/TypeScript(使用Actor.init() / Actor.exit())、Python(异步上下文管理器)以及通过CLI包装器的任何语言。提供结构化工作流:使用apify init搭建脚手架,应用SDK封装,配置输入/输出模式,通过apify run进行本地测试,然后使用apify push进行部署。包含输入和输出模式验证、Docker容器化以及可选的按事件付费...
apify-content-analytics
apify
通过Apify Actors实现多平台内容分析,支持Instagram、Facebook、YouTube和TikTok。涵盖17+个专业Actors,覆盖所有四个平台的帖子、Reels、故事、评论、话题标签、粉丝和广告。使用mcpc CLI动态获取Actor模式,以确定所需输入和可用输出字段。结果以三种格式输出:快速聊天显示、CSV导出或JSON导出,并可自定义结果数量。需要在.env文件中配置Apify令牌,并安装Node.js 20.6+...
apify-ecommerce
apify
从50多个电商平台提取产品数据、价格、评论和卖家信息。三种工作模式:产品与定价(价格追踪、竞品分析)、客户评论(情感分析、质量问题)和卖家情报(通过Google Shopping发现供应商)。支持亚马逊(20多个地区)、沃尔玛、eBay、宜家、好市多及欧洲零售商;可通过产品链接、分类链接或关键词搜索输入。可选AI驱动分析生成价格洞察...
apify-generate-output-schema
apify
通过分析Apify Actor的源代码生成输出模式(dataset_schema.json、output_schema.json、key_value_store_schema.json)。在以下情况下使用…
apify-influencer-discovery
apify
使用Apify Actors在Instagram、Facebook、YouTube和TikTok上发现并评估网红。将发现请求路由至15+个专业Actors,涵盖所有主流平台的个人资料抓取、标签搜索、互动分析和细分领域发现。通过mcpc动态获取Actor模式,在执行前确定所需输入和可用输出字段。支持三种导出模式:内联聊天显示、CSV或JSON文件输出,并可自定义结果数量...
apify-ultimate-scraper
apify
自动化网页抓取工具,为55多个平台(包括Instagram、TikTok、YouTube、Facebook、Google Maps等)选择最优Actor。涵盖8大主流平台的55多个预配置Actor,并提供针对特定用例的选择指导(潜在客户生成、网红发现、品牌监控、竞争对手分析、趋势研究)。支持三种输出格式:快速聊天显示、CSV导出或JSON导出,可自定义结果数量限制。包含多Actor工作流模式,适用于复杂...