Agent Library · MCP IO

Actor & Scraper

A complete Actor & Scraper agent library — 10 column prompts, each written out in full with its questionnaire and the fields it produces, ready to run in your own workspace.

  • 10column prompts
  • 6part instructions
  • 1time payment
  • ∞yours to keep

The Notion template

What you duplicate into your workspace

One page with the library’s database — a column per prompt, in the order they run — and every column prompt written out underneath it, ready for Notion AI autofill.

Actor & Scraper open in Notion on a desktop: the Prompt Library database with its column prompts as columns
The library database · Prompt Library, Generate and Gallery views
The column prompts section of Actor & Scraper: one toggle per column, the first open on its six parts and fields
Column prompts · six parts, a questionnaire and the fields each writes
Actor & Scraper on a phone, beside the Duplicate button
On mobile · duplicate it and run it anywhere

What’s inside

10 column prompts, in the order they run

Each one is a column of the library: it asks its questions, writes its fields, and hands them to the prompts after it.

  1. 01Agent Use Cases
  2. 02Actor Selection
  3. 03Tool Contract Design
  4. 04Run Lifecycle Handling
  5. 05Result Shaping Rules
  6. 06Authentication and Tokens
  7. 07Usage Guardrails
  8. 08Tool Call Evaluation
  9. 09Tool Usage Measures
  10. 10Server Listing and Docs

A look inside

Prompt 01 · Agent Use Cases

Introduction

Analyze all of the Company Information on the Page and in the Page Properties to generate an Agent Use Cases as a structured question set, that clearly establishes the agent use cases component of the Actor & Scraper Prompt Library. Purpose and value: Establishes what agents will ask scraping tools to do, so only actors that serve real tasks are exposed.

  • Key terms
  • Context
  • Creation
  • Framework
  • Constraints
  • Questionnaire
  • Fields it writes

Fields this prompt writes

  • {{01_Agent_Use_Cases}}
  • {{AGENT_TASKS}}
  • {{LIVE_DATA_NEEDS}}
  • {{AVAILABLE_AGENT_INPUTS}}
  • {{EXPECTED_ANSWERS}}

The prompts after it read these, which is what makes the library a chain rather than a list.

How it works

From checkout to first output in minutes

  1. 1

    Buy once

    Pay through Stripe. No account, no subscription, no seat.

  2. 2

    Duplicate in Notion

    The confirmation page gives you the Notion template: press Duplicate and it lands in your own workspace.

  3. 3

    Run the chain

    Paste each column prompt in order, answer its questionnaire, and it writes the fields the next one reads.

Where it runs

The loop it is part of, and its own workflow

Every Agent Library sits on one stage of the Intelligent Operations loop. Inside, its column prompts run in five phases, with a person reviewing the output before any of it ships, and monitoring that starts the next run.

Review the briefReview the outputReview the results01Know02Research03Plan04Create05Distribute06Convert07Retain08MeasureIntelligent OperationsOperateAgents, data and automation
The hub · OperateAgents, data and automationThe hub runs every stage: it moves fields between libraries, schedules runs and watches for change.

Takes from

Hands to

Phases

  1. Discover2 prompts
  2. Define1 prompt
  3. Build3 prompts
  4. Validate2 prompts
  5. Activate2 prompts
New brief or recordA row in the libraryQuestionnaire intake10 columns to answerInputscomplete?Discover · 201 Agent Use Cases02 Actor SelectionDefine · 103 Tool Contract DesignBuild · 304 Run Lifecycle Handling05 Result Shaping Rules06 Authentication and TokensYesNoPull fields fromWeb Scraper Actor Specif…Validate · 2A person reviews the output07 Usage Guardrails08 Tool Call EvaluationReadyto ship?Not readyApprovedActivate · 209 Tool Usage Measures10 Server Listing and DocsHand off toData Ingestion & Enrichm…Input-to-Outcome MappingMonitor resultsFields, outcomes, driftChangedetected?Yes · new runNoKeep monitoring

What you get

  • 10 column prompts, every one written out in full.
  • Six-part instructions: introduction, key terms, context, creation, framework, constraints.
  • The questionnaire each prompt asks before it writes.
  • The {{FIELDS}} each prompt produces, so the library runs as a chain.
  • A Notion template you duplicate into your workspace: the library as a database, a column per prompt.
  • The whole library as a Markdown file you keep: no login, no seat, no expiry.

About this library

Actor & Scraper is one of the Agent Libraries in MCP IO. The design for exposing web scraping actors to AI agents as Model Context Protocol tools: which actors, their tool contracts, run handling, result shaping and guardrails. It holds 10 column prompts, in the order they are meant to run, each one a six-part instruction — introduction, key terms, context, creation, framework and constraints — with the questions it asks first and the fields it writes for the prompts after it.

Questions

How is it delivered?

Instantly. After payment you get the Notion template to duplicate into your workspace, and a download link for the complete library as a Markdown file.

What do I run it in?

Any model or workspace that takes a prompt — paste a column prompt, answer its questionnaire, and it writes the fields the next one reads.

What is the licence?

Single-workspace licence. For your own use; not for resale or redistribution.

Is this a subscription?

No. One payment, and the file is yours to keep and edit.

Agent Library

Take Actor & Scraper into your own workspace

Actor & Scraper$14 · one-time