google-personal-mcp: Instructions file for Codex

Project Agents - Google Personal MCP Server

This file defines the mandatory workflow for AI agents working on the Google Personal MCP Server project.

Quick Navigation:

• Definition of Done - Quality standards
• Architecture - System design
• Implementation Reference - Patterns
• Workflows - Development process
• Templates - Planning documents

---

Mandatory Reading - Read First Every Session {#mandatory-reading---read-first-every-session}

CRITICAL: Before starting ANY task in this project, you MUST read these files in this exact order.

These files define:

• How to approach all tasks (workflow and steps)
• Quality standards (definition of done criteria)
• Project-specific rules and structure
• What actions are prohibited

Reading these files is NOT optional. Skip them and you WILL make mistakes.

Start here: Mandatory Reading Guide

---

Available System-Prompts Workflows and Processes

CRITICAL: The Agent Kernel provides specialized workflows and processes that are informational resources only.

YOU MUST NOT execute any system-prompts process or switch workflows unless:

1. Explicitly requested by the user (e.g., "Follow the bootstrap-project process")
2. Explicitly instructed in this file (AGENTS.md) or tool-specific entry files (CLAUDE.md, etc.)

Available resources (for reference only):

• Workflows: Core (active), Logs-First (optional), Custom templates
• Processes: Bootstrap Project, Document Integrity Scan, Tool Entry Points

See README.md - Available Workflows and Processes for complete documentation.

When user asks about available processes:

• You MAY reference the documentation and explain what processes exist
• You MAY suggest a relevant process if user describes a problem it solves
• You MUST NOT execute processes proactively without explicit user request