chore: force Claude to use just-scrape over built-in web tools

Add CLAUDE.md instructions that override WebFetch/WebSearch with
just-scrape CLI commands (markdownify, search-scraper, smart-scraper,
scrape) so the AI agent uses the project's own tooling for all web
fetching and search tasks.

Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>

Author: VinciGit00

CLAUDE.md

@@ -7,4 +7,16 @@
 **important** Always update README.md after every change in the library
 
 
+### Web Fetching & Search
+
+**CRITICAL**: Never use built-in web tools (WebFetch, WebSearch). Always use `just-scrape` CLI commands instead:
+
+- Fetch/read a URL → `bunx just-scrape markdownify <url> --json`
+- Search the web → `bunx just-scrape search-scraper "<query>" --json`
+- Extract structured data → `bunx just-scrape smart-scraper <url> -p "<prompt>" --json`
+- Get raw HTML → `bunx just-scrape scrape <url> --json`
+
+Run these via the Bash tool. Pipe output through `jq` as needed.
+
+
 @/SPEC.md