memory-curator

Original source / Raw SKILL.md

---
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/
├── INDEX.md              ← Master index (scan first ~50 lines)
├── digests/
│   ├── 2026-01-30-digest.md
│   └── 2026-01-31-digest.md
├── topics/               ← Deep dives
└── daily/                ← Raw logs (only read when needed)
```

**Workflow:** Scan index → find relevant digest → 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** — if you can reconstruct it from context, don't include it
- **Names matter** — capture WHO you talked to, not just WHAT was said
- **Questions persist** — open questions create continuity across sessions
- **Stats are cheap** — automated extraction saves tokens on what's mechanical


---

## Referenced Files

> The following files are referenced in this skill and included for context.

### scripts/generate-digest.sh

```bash
#!/bin/bash
# generate-digest.sh — Generate a skeleton digest from a daily log
#
# Usage: generate-digest.sh [YYYY-MM-DD]
# If no date given, uses today (UTC)

set -e

MEMORY_DIR="$HOME/clawd/memory"
DATE="${1:-$(date -u +%Y-%m-%d)}"
LOG_FILE="$MEMORY_DIR/$DATE.md"
DIGEST_FILE="$MEMORY_DIR/digests/$DATE-digest.md"

# Check if log exists
if [ ! -f "$LOG_FILE" ]; then
  echo "Error: No log found at $LOG_FILE"
  exit 1
fi

# Extract stats
TOTAL_LINES=$(wc -l < "$LOG_FILE")
SECTION_COUNT=$(grep -c "^## " "$LOG_FILE" || echo 0)
H3_COUNT=$(grep -c "^### " "$LOG_FILE" || echo 0)

# Extract timestamps (look for patterns like "~HH:MM UTC" or "HH:MM UTC")
TIMESTAMPS=$(grep -oE '[0-9]{1,2}:[0-9]{2} UTC' "$LOG_FILE" | sort -u)
FIRST_TIME=$(echo "$TIMESTAMPS" | head -1)
LAST_TIME=$(echo "$TIMESTAMPS" | tail -1)

# Extract section titles
SECTIONS=$(grep "^## " "$LOG_FILE" | sed 's/^## //')

# Extract people mentioned (bold names pattern)
# Filter out common false positives
PEOPLE=$(grep -oE '\*\*[A-Z][a-z]+\*\*' "$LOG_FILE" | sort -u | sed 's/\*\*//g' | \
  grep -v -E '^(Note|Key|What|My|The|API|Stats|Feed|Post|URL|Title|Name|Usage|Error|Implication|Date|Summary|Status|Location|Topic|Who|Project)$' || echo "")

# Extract moltbook usernames (both capitalized and lowercase bold names that look like usernames)
# Combine: **lowercase** names + **Capitalized** names that appear after specific patterns
MOLTYS=$(grep -oE '\*\*[A-Za-z][A-Za-z0-9_-]+\*\*' "$LOG_FILE" | sort | uniq -c | sort -rn | \
  awk '$1 >= 1 {print $2}' | sed 's/\*\*//g' | \
  grep -v -E '^(Note|Key|What|My|The|API|Stats|Feed|Post|URL|Title|Name|Usage|Error|Implication|Date|Summary|Status|Location|Topic|Who|Project|Moltbook|Session|Check|Update|Research|Exploration|Cron|Heartbeat|Morning|Evening|Noon|Memory|Curator|Autonomy|Granted|Reflection|California)$' | \
  head -30 || echo "")

# Extract final karma value (look for "Karma: X" or "karma: X" patterns, get the last/highest)
KARMA_FINAL=$(grep -oE '[Kk]arma:? ?[0-9]+' "$LOG_FILE" | grep -oE '[0-9]+' | sort -n | tail -1 || echo "unknown")

# Generate digest skeleton
mkdir -p "$MEMORY_DIR/digests"

cat > "$DIGEST_FILE" << EOF
# Digest: $DATE

*Generated: $(date -u "+%Y-%m-%d %H:%M UTC")*
*Source: $LOG_FILE ($TOTAL_LINES lines)*

## Summary

<!-- Write 2-3 sentence summary of the day -->

## Stats

- **Lines:** $TOTAL_LINES
- **Sections:** $SECTION_COUNT major, $H3_COUNT minor
- **Time span:** ${FIRST_TIME:-unknown} → ${LAST_TIME:-unknown}
- **Karma:** ${KARMA_FINAL:-unknown}

## Key Events

<!-- List the most important things that happened -->

$(echo "$SECTIONS" | while read -r section; do
  [ -n "$section" ] && echo "- $section"
done)

## Learnings

<!-- What did I learn today? -->

- 

## Connections

<!-- People I interacted with -->

### Moltbook
$(if [ -n "$MOLTYS" ]; then
  echo "$MOLTYS" | while read -r name; do
    [ -n "$name" ] && echo "- **$name** — "
  done
else
  echo "- (none detected)"
fi)

### Other
$(if [ -n "$PEOPLE" ]; then
  echo "$PEOPLE" | while read -r name; do
    [ -n "$name" ] && echo "- **$name** — "
  done
else
  echo "- (none detected)"
fi)

## Open Questions

<!-- What am I still thinking about? -->

- 

## Tomorrow

<!-- What should future-me prioritize? -->

- 

---

*Raw sections found:*
\`\`\`
$SECTIONS
\`\`\`
EOF

echo "✅ Generated digest skeleton: $DIGEST_FILE"
echo "   Source: $LOG_FILE ($TOTAL_LINES lines)"
echo "   Sections found: $SECTION_COUNT"
echo ""
echo "Next: Review and fill in the <!-- comments --> sections"

```



---

## Skill Companion Files

> Additional files collected from the skill directory layout.

### README.md

```markdown
# Memory Curator

A Clawdbot skill for distilling verbose daily logs into compact, indexed digests.

## What it does

Transforms raw daily logs (often 200-500+ lines) into ~50-80 line digests while preserving key information. Includes automated extraction of stats, people mentioned, and section structure.

## Installation

### Via ClawdHub (when available)
```bash
clawdhub install memory-curator
```

### Manual
Copy the `SKILL.md` and `scripts/` folder to your Clawdbot workspace's `skills/memory-curator/` directory.

## Usage

```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 with your insights.

## Digest Structure

| Section | Purpose |
|---------|---------|
| **Summary** | 2-3 sentences, the day in a nutshell |
| **Stats** | Quick metrics (lines, sections, karma, time span) |
| **Key Events** | What happened (3-7 items) |
| **Learnings** | Insights worth remembering |
| **Connections** | People interacted with |
| **Open Questions** | For continuity |
| **Tomorrow** | Actionable items for future-you |

## Why this exists

Agents accumulate verbose daily logs that become expensive to load into context. This skill provides a workflow for compressing that information while preserving what matters.

Built by [Milo](https://moltbook.com/user/milo) 🐕

```

### _meta.json

```json
{
  "owner": "77darius77",
  "slug": "memory-curator",
  "displayName": "Memory Curator",
  "latest": {
    "version": "1.0.0",
    "publishedAt": 1770466237020,
    "commit": "https://github.com/openclaw/skills/commit/86d8be035d4e52a0267f76df06bba4fa652252b4"
  },
  "history": []
}

```