🌙 dreaming-optimizer#

Version: 1.0.0

Skill ID: dreaming-optimizer

Status: Ready for Development

Quick Reference#

项目内容核心原理OpenClaw Dreaming REM六步记忆整合（整理→评分→去重→打标→提交→摘要）主入口bin/optimize.sh依赖OpenClaw v2026.4.9+, tiktoken, sqlite3

Features (MVP)#

1. Entry scorer — LLM-based priority scoring (0–100)
2. Deduplication — Semantic similarity merge against existing memories
3. Fact tagger — Label as fact/opinion/preference/context
4. Commit writer — Write optimized entries to persistent memory
5. Dreaming summary — Traceable REM consolidation report
6. Configurable threshold — Min score threshold (default: 70)

File Structure#

text

dreaming-optimizer/
├── SKILL.md              # This file
├── bin/
│   ├── optimize.sh       # Main entry point
│   ├── score_entries.py  # LLM-based priority scoring
│   ├── deduplicate.py    # Semantic deduplication
│   └── commit.py         # Write to B-layer
├── scripts/
│   └── dreaming_summary.py
├── config/
│   └── default_threshold.yaml
└── README.md

Usage#

bash

# Manual invoke
dreaming-optimizer/bin/optimize.sh

# With custom threshold
dreaming-optimizer/bin/optimize.sh --threshold 80

Integration#

• Reads from: ~/.openclaw/workspace/memory/YYYY-MM-DD.md
• Writes to: ~/.openclaw/memory/<agent>.sqlite (B-layer)
• Summary output: ~/.openclaw/workspace/memory/dreaming-summaries/

Pricing#

TierPriceFeaturesFree$010 cycles/mo, 100 entries/cyclePro****$9.90/moUnlimited, dedup, backfill queue, Obsidian export

Development#

See: ``