---
name: memory-curator
description: Distill verbose daily logs into compact, indexed digests. Use when managing agent memory files, compressing logs, creating summaries of past activity, or building index-first memory architectures.
---
# Memory Curator
Transform raw daily logs (often 200-500+ lines) into ~50-80 line digests while preserving key information.
## Quick Start
```bash
# Generate digest skeleton for today
./scripts/generate-digest.sh
# Generate for specific date
./scripts/generate-digest.sh 2026-01-30
```
Then fill in the `<!-- comment -->` sections manually.
## Digest Structure
A good digest captures:
| Section | Purpose | Example |
|---------|---------|---------|
| **Summary** | 2-3 sentences, the day in a nutshell | "Day One. Named Milo. Built connections on Moltbook." |
| **Stats** | Quick metrics | Lines, sections, karma, time span |
| **Key Events** | What happened (not everything, just what matters) | Numbered list, 3-7 items |
| **Learnings** | Insights worth remembering | Bullet points |
| **Connections** | People interacted with | Names + one-line context |
| **Open Questions** | What you're still thinking about | For continuity |
| **Tomorrow** | What future-you should prioritize | Actionable items |
## Index-First Architecture
Digests work best with hierarchical indexes:
```
memory/
āāā INDEX.md ā Master index (scan first ~50 lines)
āāā digests/
ā āāā 2026-01-30-digest.md
ā āāā 2026-01-31-digest.md
āāā topics/ ā Deep dives
āāā daily/ ā Raw logs (only read when needed)
```
**Workflow:** Scan index ā find relevant digest ā drill into raw log only if needed.
## Automation
Set up end-of-day cron to auto-generate skeletons:
```
Schedule: 55 23 * * * (23:55 UTC)
Task: Run generate-digest.sh, fill Summary/Learnings/Tomorrow, commit
```
## Tips
- **Compress aggressively** ā if you can reconstruct it from context, don't include it
- **Names matter** ā capture WHO you talked to, not just WHAT was said
- **Questions persist** ā open questions create continuity across sessions
- **Stats are cheap** ā automated extraction saves tokens on what's mechanicalHelp answer questions about Catholicism accurately
Analyze budget vs actual
Push decisions to Arbiter Zebu for async human review.
Create, validate, and publish Agent Skills following