skills.rest
sanjanb
sanjanbCOMMUNITY
Explore all skills by this author →
GitHub

scrape

Extract structured data from web pages via browser primitives and return JSON.

1|Updated Jul 23, 2026
Search Tags:#data extraction#browser automation#json#web scraping#read-only#intent routing#gstack
One-click install
npx skills add https://github.com/sanjanb/my-agent-harness --skill scrape-sanjanb
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: scrape
Source: https://github.com/sanjanb/my-agent-harness/tree/main/skills/scrape
Command: npx skills add https://github.com/sanjanb/my-agent-harness --skill scrape-sanjanb

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Manually pulling data off web pages is slow and error-prone, and one-off scraping scripts are hard to reuse. This Skill turns page-scraping intents into a repeatable flow: the first request prototypes the extraction through browser primitives and returns JSON, and subsequent matching requests route to a codified browser-skill that responds in about 200ms. ## Core Features & Use Cases - Intent-based scraping: Triggered by phrases like "scrape this page", "get data from", "pull from", "extract from", or "what's on" a page. - Self-codifying flow: First call on a new intent prototypes the extraction via $B primitives; later calls on a matching intent reuse the codified browser-skill for fast responses. - Read-only safety: Restricted to data extraction; mutating flows such as form fills, clicks, and submissions are delegated to the /automate skill instead. - Use Case: Ask "what's on this pricing page" and receive structured JSON of the page's data, then repeat the same request later and get the answer in milliseconds from the codified skill. ## Quick Start Ask the agent to scrape the data from a specific web page URL and return the result as JSON. Note: the Skill runs a gstack environment preamble (session detection, config, telemetry opt-in prompts) before executing the scrape workflow.

Frequently Asked Questions about scrape

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I scrape data from a web page with this skill?▼

Invoke it with a natural phrase like "scrape this page" or "get data from" followed by the target page. The first call prototypes the extraction through browser primitives and returns the result as JSON.

How does the scrape skill make repeat requests faster?▼

The first call on a new intent prototypes the flow via $B primitives. Subsequent calls matching that intent route to a codified browser-skill, returning results in roughly 200 milliseconds.

Can the scrape skill fill forms or click buttons on a page?▼

No. The skill is strictly read-only and limited to data extraction. Mutating flows such as form fills, clicks, and submissions must use the separate /automate skill instead.

What triggers the scrape skill automatically?▼

It activates on phrases like "scrape", "get data from", "pull", "extract from", or "what's on" a page. If gstack proactive mode is disabled, the agent asks before invoking it.

Why does the scrape skill run a long preamble before scraping?▼

The preamble detects session type, branch, config, telemetry consent, and upgrade state so the workflow behaves correctly in interactive, headless, and spawned sessions. It is a one-time setup cost per invocation.

Related Skills

Explore More Software Engineering Skills→
OpenCovenOpenCoven
20

tinyfish-web-agent

Extract JSON data from web pages via the run-sse endpoint.

Official
Advanced
KaranKathur06KaranKathur06

data-pipeline

Build scheduled ETL workflows that extract, transform, and load data into analytics destinations.

Community
Intermediate
gganbukim1gganbukim1

data-scraper-agent

Scrape public web sources on a schedule and enrich items with Gemini summaries.

Community
Advanced
TwoTonedTwoToned

scrape

Prototype web scraping flows with shell-like primitives and return JSON.

Official
Advanced