Does the agent account for the Facebook learning phase?

Yes, the skill tracks the creation date of objects and applies a 48-hour 'softening' period where optimizations are less aggressive to allow the algorithm to stabilize.

How does the Health Score (HS) work in Ads Optimizer?

The Health Score is a value between -100 and +100 calculated by weighing the CPL gap against targets, 3-day performance trends, diagnostic metrics like CTR/CPM, and statistical volume factors.

How does it handle budget redistribution?

When an ad set is paused or its budget is reduced, the 'freed budget' is automatically reallocated to high-performing ad sets or used to test new creatives from your library.

Can this skill manage multiple business units within one account?

Yes, it maps campaigns to specific directions (e.g., different product lines) and manages individual budgets and CPL targets for each independently.

What is an 'Ad-Eater' in this context?

An Ad-Eater is a specific creative within an ad set that consumes a significant portion of the budget (usually over 50%) without generating leads or meeting CPL targets.

Ads Optimizer for Facebook

Name: Ads Optimizer for Facebook
Author: dengineproblem

bydengineproblem

0•

Marketing Automation

Optimizes Facebook Ads performance by analyzing metrics, managing budgets across business directions, and detecting underperforming creative assets.

Ads Optimizer is a specialized agent for Facebook Ads Manager designed to automate the complex task of campaign optimization. It calculates a multi-dimensional Health Score for ad sets, identifies "ad-eaters" that drain budget without conversions, and ensures account spending aligns with strict business objectives. By integrating historical performance data with real-time MCP tool insights, it provides granular recommendations for scaling successful ads, pausing losers, and rotating new creatives to maintain a healthy cost-per-lead (CPL) across multiple product lines or services.

Key Features

01Granular Ad-Eater detection to identify and pause specific underperforming ads within active ad sets.

02Historical action tracking to respect 48-hour learning phases and prevent redundant adjustments.

030 GitHub stars

04Automated budget balancing and redistribution across diverse business directions and account caps.

05Multi-metric Health Score (HS) calculation based on CPL gap, trends, diagnostics, and volume factors.

06Intelligent creative rotation suggestions based on unused assets and available testing budget.

Use Cases

01Systematically scaling high-performing ad sets while ensuring overall account budget compliance.

02Automating daily budget management to maintain target CPL for high-volume lead generation campaigns.

03Auditing complex Facebook Ads accounts to find hidden budget waste at the individual creative level.

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add dengineproblem/agents-monorepo ads-optimizer

For use in Claude.ai and ChatGPT

Download Skill