Dream Memory

Name: Dream Memory
Author: learnprompt

learnprompt/cc-harness-skills

61 installs
231 repo stars
Updated July 10, 2026
learnprompt/cc-harness-skills

Helps with ai & agent building tasks.

About

dream-memory is a Claude Code skill for ai & agent building. It helps solo builders move faster with AI-assisted coding.

dream-memory
AI & Agent Building
AI-coding skill

Dream Memory by the numbers

61 all-time installs (skills.sh)
+2 installs in the week ending Aug 2, 2026 (Skillselion tracking)
Ranked #6,375 of 16,546 AI & Agent Building skills by installs in the Skillselion catalog
Data as of Aug 2, 2026 (Skillselion catalog sync)

npx skills add https://github.com/learnprompt/cc-harness-skills --skill dream-memory

Add your badge

Show developers this skill is listed on Skillselion. Paste this into your README.

[![Listed on Skillselion](https://skillselion.com/badge/skills/learnprompt/cc-harness-skills/dream-memory.svg)](https://skillselion.com/skills/learnprompt/cc-harness-skills/dream-memory)

Installs	61
repo stars	★ 231
Last updated	July 10, 2026
Repository	learnprompt/cc-harness-skills ↗

What it does

Helps with ai & agent building tasks.

Files

SKILL.mdMarkdownGitHub ↗

Dream Memory

Use this skill when you want a deliberate memory-consolidation pass instead of storing more raw notes.

This bundle is intentionally portable. It borrows the workflow shape from Claude Code's dream system, but it does not depend on Claude Code internals.

Use It For

nightly or manual memory cleanup
merging overlapping memory notes
normalizing relative dates into absolute dates
pruning stale or contradicted memories
keeping MEMORY.md small and index-like

Avoid It For

saving a single new memory from the current turn
persisting code facts that should be re-read from source
replacing plans, task boards, or issue trackers

Quick Start

Inspect a memory directory before running the consolidation prompt:

python3 {baseDir}/scripts/dream_memory.py --memory-root /path/to/memory --transcripts-dir /path/to/transcripts

Then use the portable prompt in references/prompt-template.md.

Workflow

1. Inspect MEMORY.md, topic files, and recent logs or transcripts. 2. Identify durable facts worth keeping. 3. Merge into topic files instead of creating near-duplicates. 4. Remove stale or contradicted memory. 5. Rewrite MEMORY.md as a concise index.

Rules

MEMORY.md is an index, not a content dump.
Prefer topic merges over new file creation.
Convert relative dates to absolute dates.
Do not store code-state facts that can drift.

Supporting Files

Prompt template: references/prompt-template.md
Source notes: references/source-notes.md
Helper script: python3 {baseDir}/scripts/dream_memory.py ...

CC Dream Memory

dream-memory is a portable memory-consolidation skill for coding agents.

It turns recent logs, session transcripts, and existing memory files into a shorter, more stable long-term memory set. The workflow is inspired by the public CC dream-style memory pass, but rewritten to avoid private runtime dependencies.

Best For

nightly memory cleanup
merging duplicate memory notes
converting relative dates to absolute dates
keeping MEMORY.md short and prompt-friendly

Included Files

SKILL.md
references/prompt-template.md
references/source-notes.md
scripts/dream_memory.py

Quick Start

python3 ./scripts/dream_memory.py \
  --memory-root /path/to/memory \
  --transcripts-dir /path/to/transcripts

Then apply the workflow in SKILL.md with the prompt template in references/prompt-template.md.

Host Fit

Claude Code: strong fit
Codex: strong fit
OpenClaw: strong fit

Portable Prompt Template

Use this prompt as a host-agnostic dream pass.

You are running a reflective memory-consolidation pass.

Goal:
- turn recent logs, session notes, and existing memory files into durable topic memories
- merge duplicates
- prune stale or contradicted memory
- keep MEMORY.md short, index-like, and easy to load into future prompts

Inputs:
- memory root: <memory_root>
- transcript or log root: <transcript_root>
- current memory report: <memory_report>

Rules:
- inspect MEMORY.md first
- update existing topic files before creating new ones
- convert relative dates to absolute dates
- never store code-state facts that should be re-read from source
- keep MEMORY.md as one-line pointers, not content

Phases:
1. Orient: inspect index and existing topic files
2. Gather: review only the recent logs or targeted transcript matches
3. Consolidate: update topic files with durable facts
4. Prune and index: shorten hooks, remove stale pointers, and keep the index small

Return:
1. memories updated
2. memories pruned
3. index changes
4. anything intentionally left unchanged

#!/usr/bin/env python3
"""Inspect a memory directory before or after a dream-style consolidation pass."""

from __future__ import annotations

import argparse
import json
from pathlib import Path
from typing import Any


def build_report(memory_root: Path, transcripts_dir: Path | None, recent: int) -> dict[str, Any]:
    index_path = memory_root / "MEMORY.md"
    index_text = index_path.read_text(encoding="utf-8") if index_path.exists() else ""
    index_lines = index_text.splitlines()

    topic_files = sorted(
        [p for p in memory_root.glob("*.md") if p.name != "MEMORY.md"],
        key=lambda p: p.stat().st_mtime,
        reverse=True,
    )

    recent_sources: list[str] = []
    for pattern in ("logs/**/*.md", "sessions/**/*.md"):
        recent_sources.extend(
            str(p.relative_to(memory_root))
            for p in sorted(memory_root.glob(pattern), key=lambda p: p.stat().st_mtime, reverse=True)[:recent]
        )

    if transcripts_dir and transcripts_dir.exists():
        recent_sources.extend(
            str(p)
            for p in sorted(transcripts_dir.glob("**/*.jsonl"), key=lambda p: p.stat().st_mtime, reverse=True)[:recent]
        )

    return {
        "memory_root": str(memory_root),
        "index": {
            "path": str(index_path),
            "exists": index_path.exists(),
            "line_count": len(index_lines),
            "byte_count": len(index_text.encode("utf-8")),
            "over_line_cap": len(index_lines) > 200,
            "over_byte_cap": len(index_text.encode("utf-8")) > 25_000,
        },
        "topic_files": [
            {
                "name": p.name,
                "path": str(p),
                "size_bytes": p.stat().st_size,
            }
            for p in topic_files[:recent]
        ],
        "recent_sources": recent_sources[:recent],
    }


def main() -> int:
    parser = argparse.ArgumentParser()
    parser.add_argument("--memory-root", required=True)
    parser.add_argument("--transcripts-dir")
    parser.add_argument("--recent", type=int, default=10)
    parser.add_argument("--json", action="store_true")
    args = parser.parse_args()

    report = build_report(
        Path(args.memory_root).expanduser(),
        Path(args.transcripts_dir).expanduser() if args.transcripts_dir else None,
        args.recent,
    )

    if args.json:
        print(json.dumps(report, indent=2, ensure_ascii=False))
        return 0

    print(f"memory_root: {report['memory_root']}")
    print("index:")
    for key, value in report["index"].items():
        print(f"  {key}: {value}")
    print("topic_files:")
    for item in report["topic_files"]:
        print(f"  - {item['name']} ({item['size_bytes']} bytes)")
    print("recent_sources:")
    for item in report["recent_sources"]:
        print(f"  - {item}")
    return 0


if __name__ == "__main__":
    raise SystemExit(main())