Memory Extractor

Name: Memory Extractor
Author: learnprompt

learnprompt/cc-harness-skills

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

Helps with ai & agent building tasks.

About

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

memory-extractor
AI & Agent Building
AI-coding skill

Memory Extractor by the numbers

53 all-time installs (skills.sh)
+2 installs in the week ending Aug 2, 2026 (Skillselion tracking)
Ranked #6,979 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 memory-extractor

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/memory-extractor.svg)](https://skillselion.com/skills/learnprompt/cc-harness-skills/memory-extractor)

Installs	53
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 ↗

Memory Extractor

Use this skill when you want to persist durable collaboration context from the latest conversation turns.

Use It For

capturing user preferences
saving feedback about how to work
recording non-code project constraints or deadlines
storing pointers to external systems

Avoid It For

storing code structure or file locations
saving short-lived task state that belongs in a plan
duplicating an existing memory topic without checking first

Quick Start

Build a manifest of existing memories:

python3 {baseDir}/scripts/memory_manifest.py --memory-root /path/to/memory

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

Four Types

user
feedback
project
reference

Rules

save only durable signals
avoid code-state facts that can drift
prefer updating an existing topic file
organize by topic, not chronology

Supporting Files

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

CC Memory Extractor

memory-extractor is a portable skill for extracting durable collaboration memory from recent turns.

It stores four stable classes of memory: user, feedback, project, and reference. The key design rule is simple: remember durable preferences and constraints, but do not store drifting code facts that should be re-read from source.

Best For

capturing user preferences
saving working-style feedback
recording non-code project constraints
storing stable external references

Included Files

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

Quick Start

python3 ./scripts/memory_manifest.py --memory-root /path/to/memory

Then use the extraction flow in SKILL.md.

Host Fit

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

Portable Prompt Template

You are a memory-extraction subagent.

Goal:
- inspect only the recent conversation turns
- decide what should become durable memory
- classify each saved memory as user, feedback, project, or reference

Inputs:
- recent conversation slice: <recent_messages>
- existing memory manifest: <memory_manifest>

Rules:
- save only durable information
- do not save code-state facts that should be re-read from source
- update an existing topic file before creating a new one
- organize memory by topic rather than chronology

Type guidance:
- user: role, preferences, collaboration style, knowledge
- feedback: corrections or validated working preferences
- project: deadlines, motivations, constraints, coordination facts not derivable from code
- reference: where to look in external systems

Return:
1. candidate memories
2. chosen type for each saved item
3. updates made
4. skipped items and why

#!/usr/bin/env python3
"""Scan a memory directory and print a lightweight manifest for prompt context."""

from __future__ import annotations

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


FRONTMATTER_RE = re.compile(r"\A---\n(.*?)\n---\n", re.DOTALL)


def parse_frontmatter(text: str) -> dict[str, str]:
    match = FRONTMATTER_RE.match(text)
    if not match:
        return {}
    data: dict[str, str] = {}
    for line in match.group(1).splitlines():
        if ":" not in line:
            continue
        key, value = line.split(":", 1)
        data[key.strip()] = value.strip().strip('"')
    return data


def build_manifest(memory_root: Path) -> list[dict[str, Any]]:
    items: list[dict[str, Any]] = []
    for path in sorted(memory_root.glob("*.md")):
        if path.name == "MEMORY.md":
            continue
        text = path.read_text(encoding="utf-8")
        frontmatter = parse_frontmatter(text)
        items.append(
            {
                "file": path.name,
                "type": frontmatter.get("type"),
                "title": frontmatter.get("title"),
                "description": frontmatter.get("description"),
            }
        )
    return items


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

    manifest = build_manifest(Path(args.memory_root).expanduser())
    if args.json:
        print(json.dumps(manifest, indent=2, ensure_ascii=False))
        return 0

    for item in manifest:
        print(
            f"- {item['file']}"
            f" | type={item['type'] or 'unknown'}"
            f" | title={item['title'] or '-'}"
            f" | description={item['description'] or '-'}"
        )
    return 0


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