mutation-safety

Enforce two-step approval for Google Ads and Sheets mutations.

58|9|Updated Apr 7, 2026
One-click install
npx skills add https://github.com/fourteenwm/ppc-ai-skills --skill mutation-safety
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: mutation-safety
Source: https://github.com/fourteenwm/ppc-ai-skills/tree/main/mutation-safety
Command: npx skills add https://github.com/fourteenwm/ppc-ai-skills --skill mutation-safety

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Two-step approval system to prevent accidental mutations to Google Ads accounts and destructive Google Sheets writes, reducing the risk of costly live changes.

Core Features & Use Cases

  • Dry-run previews that show current vs. proposed changes and count of entities affected.
  • Explicit user approval before any execution to ensure intentional mutations.
  • Exact-match and scope verification to prevent mis-targeted edits across accounts.
  • Works for Google Ads API mutations and destructive Google Sheets writes with strong guardrails.

Quick Start

Copy SKILL.md into your Claude Code project at .claude/skills/mutation-safety and Claude will automatically enforce the two-step approval process for all mutations.

Frequently Asked Questions about mutation-safety

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

FAQPage Schema
How do I prevent accidental mutations to my Google Ads account when using AI agents?

A two-step approval workflow prevents accidental mutations to Google Ads accounts by requiring a dry run preview, explicit user confirmation, and exact-match scope verification before any execution. This ensures safe, auditable changes within managed accounts.

What is a dry run preview for Google Ads mutations and how does it work?

A dry run preview for Google Ads mutations displays current versus proposed changes and counts affected entities before execution. It acts as the first step in a two-step guardrail process, allowing you to verify exact matching and scope before explicitly approving the action.

Can I use Claude to apply guardrails against destructive Google Sheets writes?

Yes, Claude can apply guardrails against destructive Google Sheets writes. By copying the skill configuration into your Claude Code project, the agent automatically enforces a two-step approval process with dry run previews and explicit user approval for all destructive writes.

How do I set up two-step approval for Google Ads changes in Claude Code?

To set up two-step approval for Google Ads changes, copy the skill file into your Claude Code project directory at .claude/skills/mutation-safety. Claude will then automatically enforce the dry run preview, explicit approval, and separate execution workflow for all mutations.

Does the two-step mutation guardrail verify scope to prevent mis-targeted edits across accounts?

Yes, the two-step mutation guardrail verifies scope to prevent mis-targeted edits across accounts. It enforces exact matching during the execution step to ensure that mutations only affect the intended entities within your managed accounts.

What are the limitations of using automated guardrails for Google Ads API mutations?

The primary limitation of automated guardrails for Google Ads API mutations is the mandatory two-step workflow, meaning no action executes without explicit user approval. This prevents autonomous bulk changes and requires manual confirmation for every mutation.