원클릭으로
webmcp
Browser-native AI agent tools via navigator.modelContext.registerTool()
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
메뉴
Browser-native AI agent tools via navigator.modelContext.registerTool()
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
AI content usage preferences declared via Content-Signal in robots.txt
Get any page as Markdown by sending Accept: text/markdown
Browse Savta's bilingual recipe collection via REST API
Access Savta's bilingual recipe collection via Model Context Protocol
Rotate incorrectly oriented recipe scan images. Use this skill whenever a user mentions a recipe scan being sideways, upside down, rotated wrong, or needing orientation correction — including when referencing a GitHub issue about image rotation. Also trigger when the user says "rotate", "flip", or "fix orientation" in the context of recipe scans or images.
| name | webmcp |
| type | webmcp |
| version | 1.0.0 |
| description | Browser-native AI agent tools via navigator.modelContext.registerTool() |
Savta's Recipes exposes browser-native tools to AI agents via the WebMCP API. When loaded in a browser that supports navigator.modelContext, agents can search and navigate recipes without leaving the page.
Tools are registered on every recipe page using navigator.modelContext.registerTool() (one call per tool, matching the current WebMCP spec). The older provideContext({tools}) form is retained as a runtime fallback for shims that don't implement registerTool.
Search the recipe collection by name, ingredient, or tag.
Input schema:
{ "query": { "type": "string" } }
Returns: Up to 10 matching recipes with slug, titleEn, titleHe, tags, and url.
List every recipe with names, tags, and page URLs.
Input schema: {}
Returns: All recipes with slug, titleEn, titleHe, tags, and url.
Navigate the browser to a specific recipe page.
Input schema:
{ "slug": { "type": "string" } }
Navigate to the search page, optionally pre-filled with a query.
Input schema:
{ "query": { "type": "string" } }