Can Claude ever access production systems with this skill enabled?

No. Interactions with production databases, APIs, or credentials are strictly prohibited with no exceptions to ensure maximum system safety.

What happens if Claude needs to perform a destructive Git action?

For actions like force pushing or hard resets, Claude must ask for and receive explicit consent from the user within the current active session, regardless of what the issue description says.

Does this skill block all infrastructure changes?

No, it only blocks changes that have not been explicitly authorized within a GitHub issue. Once authorized, Claude can proceed with the specific requested change.

How does it handle security-related code changes?

Any modifications to authentication or authorization systems are automatically flagged for mandatory human review before they can be merged.

Maverick Scope Boundaries

Name: Maverick Scope Boundaries
Author: thermiteau

bythermiteau

0•

安全与测试

Enforces strict safety guardrails and authorization protocols to prevent unauthorized infrastructure, security, or production system changes.

The Maverick Scope Boundaries skill establishes essential operational limits for Claude Code, ensuring that autonomous actions do not compromise system integrity or security. It defines clear 'no-go' zones including CI/CD pipelines, authentication logic, and destructive Git operations, requiring explicit human authorization or session-specific consent before proceeding with high-risk tasks. By implementing a 'When in doubt, don't' principle, this skill mitigates the risks associated with AI-driven development in sensitive environments.

主要功能

01Zero-exception prohibition of production system interactions

02Standardized escalation templates for requesting human intervention

030 GitHub stars

04Strict infrastructure protection for CI/CD, Docker, and IaC configurations

05Prevention of destructive Git commands like force pushes and hard resets

06Security guardrails for authentication, authorization, and API key management

使用场景

01Ensuring security-sensitive auth middleware changes are flagged for manual review

02Preventing accidental overwriting of remote branches during automated refactoring

03Protecting production environments from unauthorized AI-driven database or API access

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add thermiteau/maverick-private mav-scope-boundaries

For use in Claude.ai and ChatGPT

主要功能

01Zero-exception prohibition of production system interactions

02Standardized escalation templates for requesting human intervention

030 GitHub stars

04Strict infrastructure protection for CI/CD, Docker, and IaC configurations

05Prevention of destructive Git commands like force pushes and hard resets

06Security guardrails for authentication, authorization, and API key management

使用场景

01Ensuring security-sensitive auth middleware changes are flagged for manual review

02Preventing accidental overwriting of remote branches during automated refactoring

03Protecting production environments from unauthorized AI-driven database or API access

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add thermiteau/maverick-private mav-scope-boundaries

For use in Claude.ai and ChatGPT