agent-config

Token-Efficient Agent Toolkit with Google Workspace Integration

Part of the 8-layer -config ecosystem by @junghan0611

한국어 | junghan0611 Profile

---

What is agent-config?

Public template and toolkit for AI agents, focusing on token efficiency and workspace automation.

Originally derived from private claude-config (Layer 4), designed for community sharing.

Position in 8-Layer Ecosystem

Layer 6: meta-config          (Agent Orchestration)
Layer 5: memex-kb, memacs     (Knowledge, Time Integration)
Layer 4: claude-config        (Private Memory) ← agent-config is public version
         agent-config         (Public Toolkit)
Layer 3: zotero-config        (Bibliography)
Layer 2: doomemacs-config     (Editor)
Layer 1: nixos-config         (OS)

---

Core Features

🎯 3-Tier Information Architecture

Tier 1: Tracking → 50 tokens (thin pointers)
Tier 2: Workspace → 300 tokens (Apps Script summaries)
Tier 3: Repository → Direct access (what you need)

🚀 Apps Script Layer

• 10-20x token savings (verified, not exaggerated)
• Complete automation: Time-driven triggers
• Custom functions: =ANALYZE(A1:A10) in Sheets
• Workspace native: Internal processing

📊 Lean Tracking

1 Repository = 1 Tracking File (1KB)

10 repos = 500 tokens (vs 10,000 traditional)

---

Quick Start

# Clone
git clone https://github.com/junghan0611/agent-config.git

# Setup Google Workspace MCP
# See docs/GOOGLE-WORKSPACE.md

# Deploy Apps Script
# See src/apps-script/mcp-chat-app/README.md

---

Philosophy

"Memory is an index, Repository is truth, Workspace is dashboard"

• Agent memory: Just pointers
• Real data: In repositories
• Details: Workspace provides
• Automation: Apps Script handles

---

Status

🚀 Active Development (2025-11-14)

Part of junghan0611's ecosystem: https://github.com/junghan0611

---

License

MIT

---

"Give agents wings with token-efficient tools"