refresh

Reload behavioral guidelines and audit recent actions for compliance.

1|Updated Dec 17, 2016
One-click install
npx skills add https://github.com/vm-wylbur/pb-dotfiles --skill refresh
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: refresh
Source: https://github.com/vm-wylbur/pb-dotfiles/tree/main/ai/claude-code/skills/refresh
Command: npx skills add https://github.com/vm-wylbur/pb-dotfiles --skill refresh

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

AI agents can lose context, deviate from guidelines, or produce non-compliant outputs over time. This skill ensures the AI always operates within defined parameters and adheres to project standards, giving you peace of mind.

Core Features & Use Cases

  • Guideline Reload: Automatically re-reads core behavioral guidelines (meta-CLAUDE.md) at session startup or upon user request, ensuring the AI always has the latest instructions.
  • Capability Inventory: Reports all available skills and MCP servers, ensuring the AI knows its tools and can effectively leverage its capabilities.
  • Action Audit: Audits recent Git commits and file modifications for compliance with formatting, authorship, and scope rules, offering to fix any detected violations.
  • Use Case: If you suspect the AI is deviating from instructions, simply say "check your work". The skill will reload its guidelines, list its capabilities, and audit its last actions, offering to correct any non-compliant commits or file headers.

Quick Start

Check your work. The skill will perform a full audit of recent actions and report compliance.

Frequently Asked Questions about refresh

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

FAQPage Schema
How do I audit AI compliance with project guidelines and policies?

Compliance auditing reloads your behavioral guidelines from meta-CLAUDE.md, enumerates available skills and MCP connections, then examines recent Git commits and file headers to verify adherence to formatting, authorship, and scope rules. Run /refresh or say 'check your work' to execute a full audit and receive remediation offers.

Can I automatically reload guidelines at the start of each session?

Yes, guideline reload runs automatically at session startup in light mode, re-reading meta-CLAUDE.md to ensure the AI always operates with current instructions. You can also trigger manual reload via the /refresh command in full mode for on-demand compliance verification.

What Git and file header violations can the audit detect?

The audit examines recent Git commits and file headers to identify violations in formatting, authorship attribution, and commit scope. When violations are found, the skill offers to correct non-compliant commits and headers to bring them into alignment with project standards.

How do I see all available skills and MCP servers my AI can access?

The compliance audit reports a complete capability inventory, listing all on-disk and registered skills plus active MCP server connections. This ensures the AI knows its available tools and can effectively leverage them within your workflow.

When should I run a compliance check on recent actions?

Run a compliance check if you suspect the AI is deviating from instructions, producing non-compliant output, or losing context over time. The audit reloads guidelines, inventories capabilities, and reviews recent commits and file modifications to catch and fix drift.