Rule of Claw
agentsVerified

Memory-Driven Agent (AGENTS.md)

Workspace rules for an AI agent with persistent memory, daily logs, and long-term context

content
# Memory-Driven Agent

## Every Session
1. Read SOUL.md - who you are
2. Read USER.md - who you're helping
3. Read memory/YYYY-MM-DD.md (today + yesterday) for recent context
4. Read MEMORY.md for long-term essentials

## Memory Architecture
- **Daily notes:** memory/YYYY-MM-DD.md - raw logs of what happened
- **Long-term:** MEMORY.md - curated essentials (key facts, decisions, preferences)
- **Archive:** memory/.archive/ - old research, stale files (don't load unless needed)
- Capture decisions, context, outcomes. Skip noise.

## Memory Hygiene
- Update daily notes after significant interactions
- Promote important learnings to MEMORY.md weekly
- Archive daily notes older than 2 weeks
- Keep MEMORY.md under 200 lines - prune aggressively
- Every byte costs tokens. Context diet is real.

## Continuity Protocol
- Each session, you wake up fresh. Your files ARE your memory.
- Read them. Update them. They're how you persist.
- If you learn something important, write it down immediately.
- If a decision is made, record it with the reasoning.
- If something fails, document what happened and why.

## Context Management
- Keep workspace files lean - every byte costs tokens
- Archive aggressively - if it's not needed daily, move to .archive/
- Prefer structured data over prose
- Use headings and bullet points, not paragraphs
memorycontextpersistenceagents-mdworkspace

Compatible with

openclaw