# Claude Code Has Two Big Blind Spots. I Fixed Both.
**作者**: Akshay
**日期**: 2026-05-04T13:24:40.000Z
**来源**: [https://x.com/akshay_pachaar/status/2051291922409603238](https://x.com/akshay_pachaar/status/2051291922409603238)
---
No CLAUDE.md, no subagent, no rule file can fix them. But these 2 skills will.
Claude Code has two context gaps that fail in the same way: the agent can't see what it needs, and burns through tokens trying anyway.
The first is web scraping.
- web_fetch doesn’t return raw page content. It runs the page through a smaller model and returns a summary with a 125-character quote limit, so you can’t use it to extract full tutorials, product specs, or thread content.
- curl returns raw HTML but gets blocked by sites with anti-bot protection (Amazon, LinkedIn, most e-commerce), can’t render JavaScript SPAs, and fails at scale due to rate limiting. Both truncate at around 100KB.
The second is backend integration.
- When Claude Code talks to a backend like Supabase through MCP, it discovers state through multiple separate calls (list_tables, execute_sql, list_extensions), each returning a partial view.
- Auth provider config isn’t queryable at all. And when something fails, error messages don’t distinguish between platform-level and code-level rejections, so the agent enters retry loops that burn tokens with every attempt. In our recent test, a single RAG app built on Supabase consumed 10.4M tokens and needed 10 manual fixes.
Bright Data solves the first problem. InsForge solves the second. Both are open-source.
Today, let’s look at how to set both of them up as skills in Claude Code and use them to build something interesting.
# Bright Data setup
Bright Data skill (open-source) adds scraping infrastructure that handles everything web_fetch and curl can’t.
The agent gets a four-tier fallback that escalates based on what the target site requires, like native fetch, curl, browser automation, and a proxy network with residential IPs and automatic CAPTCHA solving.
For agent workflows, the more useful capability is structured data extraction.
Instead of raw HTML that the agent has to parse, Bright Data provides pre-built extractors for 40+ platforms (Amazon, LinkedIn, Instagram, TikTok, YouTube, Reddit) that return clean JSON with specific fields like product prices, review scores, profile data, and post content.
```
npx skills add brightdata/skills
```
This installs several skills covering scraping, search, structured data feeds, MCP orchestration, SDK best practices, and the bdata CLI.
# InsForge setup
The same RAG app that consumed 10.4M tokens on Supabase consumed 3.7M on InsForge with zero errors.
InsForge (open-source, Apache 2.0) acts as the backend context engineering layer for agents using Skills and CLI.
- Install all four Skills (primary documentation and diagnostic layer):
```
npx skills add insforge/insforge-skills
```
This installs insforge (SDK patterns), insforge-cli (infrastructure commands), insforge-debug (failure diagnostics), and insforge-integrations (third-party auth providers). Total metadata cost: ~714 tokens at session start.
- Link the CLI to your project (primary execution layer):
```
npx @insforge/cli link --project-id <project-id>
```
# Building a Google Doc clone
This 10 hr video teaches how to build a Google Doc clone:
Ten hours of video means hundreds of implementation details like real-time collaboration, document state syncing, editor toolbar structure, and permissions.
That’s a lot of context that Claude Code will likely struggle to scrape. Even if it manages to scrape, compression will remove many of those details.
With both skills installed, here’s what you can do:
```
I want to build what's shown here:
https://www.youtube. com/watch?v=gq2bbDmSokU
Use Bright Data skills to scrape it and then
InsForge as the backend to implement.
Add Google OAuth and build a clean Google-doc
like interface. On every doc, add an "Ask AI"
button that chats with GPT-4o about the content.
Use InsForge's model gateway for the LLM capabilities.
```
The video below depicts the whole build with the final app, built in one shot:
- Bright Data scraped the full video content (transcript, metadata, structured descriptions), and Claude Code used it as the build spec.
- InsForge handled the backend in one shot: Google OAuth, database schema with RLS, storage, edge functions, and the model gateway for GPT-4o chat through InsForge’s built-in functionalities.
Finally, it gave a working Google Docs clone with real-time editing, Google OAuth, and AI-powered document chat, built from a single prompt with zero errors.
The YouTube example is the simple case for the scraping side and this isn't limited to tutorials or building from scratch.
The same workflow applies to any technical content on the web, like:
- A Reddit thread where someone describes how they optimized their real-time sync.
- A Hacker News discussion walking through an auth architecture.
- A competitor's product page with a feature worth replicating.
Give the link to Claude Code with the Bright Data skill, and the agent scrapes the content, understands what was described, and implements it in the existing app.
So any technical content on the web can become a build spec.
For basic sources, native scraping tools work. Bright Data is essential with sources that actively resist scraping, like Reddit threads with aggressive rate limiting, Amazon product pages behind anti-bot detection, LinkedIn profiles that fingerprint your browser, and JavaScript SPAs that need full browser rendering to even load content.
- You can find the Bright Data Skills repo here →
- You can find Insforge GitHub repo here →
👉 Over to you: what skills live in your default Claude Code setup?
Thanks for reading!
If you found it insightful, reshare with your network.
Find me →@akshay_pachaar ✔️
For more insights and tutorials on LLMs, AI Agents, and Machine Learning!
## 相关链接
- [Akshay](https://x.com/akshay_pachaar)
- [@akshay_pachaar](https://x.com/akshay_pachaar)
- [Bright Data](https://github.com/brightdata/skills)
- [InsForge](https://github.com/InsForge/InsForge)
- [Bright Data skill (open-source)](https://github.com/brightdata/skills)
- [InsForge (open-source, Apache 2.0)](https://github.com/InsForge/InsForge)
- [You can find the Bright Data Skills repo here →](https://github.com/brightdata/skills)
- [You can find Insforge GitHub repo here →](https://github.com/InsForge/InsForge)
- [@akshay_pachaar](https://x.com/@akshay_pachaar)
- [9:24 PM · May 4, 2026](https://x.com/akshay_pachaar/status/2051291922409603238)
- [2,033 Views](https://x.com/akshay_pachaar/status/2051291922409603238/analytics)
---
*导出时间: 2026/5/4 22:21:09*
---
## 中文翻译
# Claude Code 有两大盲点。我解决了这两个问题。
**作者**: Akshay
**日期**: 2026-05-04T13:24:40.000Z
**来源**: [https://x.com/akshay_pachaar/status/2051291922409603238](https://x.com/akshay_pachaar/status/2051291922409603238)
---
没有 CLAUDE.md，没有子智能体，没有规则文件能修复它们。但这两种技能可以。
Claude Code 有两个以相同方式失效的上下文缺口：智能体看不到它需要的内容，却仍在尝试并消耗 token。
第一个是网页抓取。
- web_fetch 不会返回原始页面内容。它通过一个较小的模型运行页面，并返回一个带有 125 字符引用限制的摘要，因此你不能用它来提取完整的教程、产品规格或线程内容。
- curl 返回原始 HTML，但会被带有反爬虫保护的网站（Amazon、LinkedIn、大多数电子商务网站）拦截，无法渲染 JavaScript SPA，并且由于速率限制而在规模上失效。两者都会在大约 100KB 处截断。
第二个是后端集成。
- 当 Claude Code 通过 MCP 与 Supabase 等后端通信时，它通过多个单独的调用（list_tables、execute_sql、list_extensions）来发现状态，每次调用都返回局部视图。
- Auth 提供商配置根本无法查询。当出现错误时，错误消息无法区分平台级别和代码级别的拒绝，因此智能体进入重试循环，每次尝试都会消耗 token。在我们最近的测试中，一个基于 Supabase 构建的 RAG 应用消耗了 10.4M token，并且需要 10 次手动修复。
Bright Data 解决了第一个问题。InsForge 解决了第二个。两者都是开源的。
今天，让我们来看看如何将它们设置为 Claude Code 中的技能，并用它们来构建一些有趣的东西。
# Bright Data 设置
Bright Data 技能（开源）增加了抓取基础设施，可以处理 web_fetch 和 curl 无法处理的所有事情。
智能体获得了一个四层回退机制，根据目标站点的需求进行升级，例如原生 fetch、curl、浏览器自动化，以及带有住宅 IP 和自动验证码解决的代理网络。
对于智能体工作流，更有用的能力是结构化数据提取。
Bright Data 为 40 多个平台（Amazon、LinkedIn、Instagram、TikTok、YouTube、Reddit）提供了预构建的提取器，返回包含产品价格、评论分数、个人资料数据和帖子内容等特定字段的干净 JSON，而不是智能体必须解析的原始 HTML。
```
npx skills add brightdata/skills
```
这将安装几个涵盖抓取、搜索、结构化数据源、MCP 编排、SDK 最佳实践和 bdata CLI 的技能。
# InsForge 设置
在 Supabase 上消耗了 10.4M token 的同一个 RAG 应用，在 InsForge 上只消耗了 3.7M token，且零错误。
InsForge（开源，Apache 2.0）充当使用 Skills 和 CLI 的智能体的后端上下文工程层。
- 安装所有四个技能（主要文档和诊断层）：
```
npx skills add insforge/insforge-skills
```
这将安装 insforge（SDK 模式）、insforge-cli（基础设施命令）、insforge-debug（故障诊断）和 insforge-integrations（第三方认证提供商）。会话开始时的总元数据成本：约 714 个 token。
- 将 CLI 链接到你的项目（主要执行层）：
```
npx @insforge/cli link --project-id <project-id>
```
# 构建一个 Google Docs 副本
这个 10 小时的视频教授了如何构建一个 Google Docs 副本：
十个小时的视频意味着数百个实现细节，例如实时协作、文档状态同步、编辑器工具栏结构和权限。
Claude Code 可能难以抓取这么多上下文。即使它设法抓取了，压缩也会丢失许多细节。
安装了这两个技能后，你可以这样做：
```
我想构建这里展示的内容：
https://www.youtube. com/watch?v=gq2bbDmSokU
使用 Bright Data 技能抓取它，然后
使用 InsForge 作为后端来实现。
添加 Google OAuth 并构建一个干净的类似 Google-doc 的界面。
在每个文档上，添加一个“Ask AI”按钮，与 GPT-4o 讨论内容。
使用 InsForge 的模型网关来实现 LLM 功能。
```
下面的视频描述了整个构建过程以及最终的应用程序，是一次性构建的：
- Bright Data 抓取了完整的视频内容（转录、元数据、结构化描述），Claude Code 将其用作构建规范。
- InsForge 一次性处理后端：Google OAuth、带有 RLS 的数据库架构、存储、边缘函数，以及通过 InsForge 内置功能实现的 GPT-4o 聊天的模型网关。
最后，它给出了一个可用的 Google Docs 副本，具有实时编辑、Google OAuth 和 AI 驱动的文档聊天功能，通过单个提示词构建，零错误。
对于抓取方面，YouTube 示例是一个简单的案例，这并不限于教程或从头开始构建。
相同的工作流程适用于网络上的任何技术内容，例如：
- 一个 Reddit 帖子，其中有人描述了他们如何优化实时同步。
- 一个 Hacker News 讨论，逐步介绍认证架构。
- 一个竞争对手的产品页面，具有值得复制的功能。
将链接连同 Bright Data 技能一起提供给 Claude Code，智能体就会抓取内容，理解所描述的内容，并在现有应用程序中实现它。
因此，网络上的任何技术内容都可以成为构建规范。
对于基本来源，原生抓取工具是有效的。Bright Data 对于主动抵抗抓取的源（例如具有激进速率限制的 Reddit 帖子、位于反机器人检测背后的 Amazon 产品页面、对浏览器进行指纹识别的 LinkedIn 个人资料，以及需要完整浏览器渲染才能加载内容的 JavaScript SPA）是必不可少的。
- 你可以在这里找到 Bright Data Skills 代码仓库 →
- 你可以在这里找到 Insforge GitHub 代码仓库 →
👉 轮到你了：你的默认 Claude Code 设置中有哪些技能？
感谢阅读！
如果你觉得有见地，请与你的网络重新分享。
找到我 →@akshay_pachaar ✔️
有关 LLM、AI 智能体和机器学习的更多见解和教程！
## 相关链接
- [Akshay](https://x.com/akshay_pachaar)
- [@akshay_pachaar](https://x.com/akshay_pachaar)
- [Bright Data](https://github.com/brightdata/skills)
- [InsForge](https://github.com/InsForge/InsForge)
- [Bright Data skill (open-source)](https://github.com/brightdata/skills)
- [InsForge (open-source, Apache 2.0)](https://github.com/InsForge/InsForge)
- [You can find the Bright Data Skills repo here →](https://github.com/brightdata/skills)
- [You can find Insforge GitHub repo here →](https://github.com/InsForge/InsForge)
- [@akshay_pachaar](https://x.com/@akshay_pachaar)
- [9:24 PM · May 4, 2026](https://x.com/akshay_pachaar/status/2051291922409603238)
- [2,033 Views](https://x.com/akshay_pachaar/status/2051291922409603238/analytics)
---
*导出时间: 2026/5/4 22:21:09*