Clawhub Memory Tiers
by @dirtyrootsstudio
Tiered memory system for OpenClaw agents. Gives agents instant context on startup with a 4-line state snapshot (L0) and 7-day rolling context (L1). Eliminate...
Step 1: Create L0.md in your agent's workspace
# L0 β [AGENT_NAME] Quick StateYou are [AGENT_NAME]. [One sentence describing role and purpose.]
Current focus: [Top priority right now β what matters most this week.]
Last run: [Date and one-line summary of last activity, or "No runs yet."]
Flags: [Critical state warnings β cron status, credit limits, blockers. "None" if clear.]
Example β a content agent:
# L0 β WRITER Quick StateYou are WRITER. Content engine for the company blog and social accounts.
Current focus: Draft 3 LinkedIn posts promoting the new pricing page.
Last run: 2026-03-15 β Generated 2 blog teasers and 1 thought leadership post.
Flags: Publishing queue has 8 posts loaded through Mar 22. No blockers.
Example β a monitoring agent:
# L0 β WATCHDOG Quick StateYou are WATCHDOG. Security monitor for the production environment.
Current focus: Track error rates after the v2.4 deploy.
Last run: 2026-03-16 β Scanned logs, no anomalies. Alert threshold at 2% error rate.
Flags: Grafana dashboard unreachable intermittently. Investigate if persists.
Step 2: Create L1.md in your agent's workspace
```markdown
L1 β [AGENT_NAME] Rolling Context
clawhub install clawhub-memory-tiers