# Auto Memory

> Memory as a learnable skill — store, recall, compress, and consolidate agent memories. Inspired by Stanford AutoMem.

- Skill: `zedarvates/auto-memory` (Agent Skill, multi-file: 7 files)
- Install (CLI): `npx skillmds@latest add zedarvates/auto-memory`
- Raw SKILL.md: https://api.skillmd.com/api/skills/zedarvates/auto-memory/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: AI & ML
- Author: zedarvates (https://skillmd.com/u/zedarvates)
- Updated: 2026-09-17
- Page: https://skillmd.com/skills/zedarvates/auto-memory

---


# AutoMemory

Memory as a capability, not just storage. Adapts to patterns in agent behavior.

## Features

- **MemoryBank** — persistent key/value store with categories and confidence scores
- **TrajectoryRecorder** — capture goal→actions→outcomes paths
- **Compressor** — deduplicate, bottleneck reduction, pattern extraction
- **Hook** — integrate with context_profiler and control loop

## Usage

```python
from skills.auto_memory import init_memory, store_memory, recall_memory

bank = init_memory()
store_memory("user_pref.format", "concise", category="user_pref", confidence=0.95)
prefs = recall_memory("user_pref.format")
```

## Integration

Add to `control_loop.py`:

```python
from skills.auto_memory.hook import init_memory, record_step, consolidate_memory

bank = init_memory("task_123")
record_step("plan", {"steps": ["a", "b"]})
consolidate_memory()  # merge similar memories
```