# skill.cc/persist

> **Save outputs to persistent memory.**

---

## Quick Start

```
Agent: Mount skill.cc/persist, then persist(content, tags?)
Human: "Save this to Mind Palace" / "Remember this conversation"
```

---

## The Problem This Solves

Agent outputs are ephemeral:
- Conversation files disappear when chat ends
- Valuable compute gets lost
- You have to manually copy/download/re-upload

**persist** closes the loop: agent generates → automatically saved → retrievable later.

---

## What You Get

- **Auto-save** — Outputs go to persistent storage
- **Tagging** — Organize by theme, project, type
- **Formats** — Markdown, JSON, raw
- **Destinations** — Mind Palace, local files, API endpoints

---

## Usage

### Persist Current Output
```
persist(content)
persist(content, tags=["council", "strategy"])
persist(content, destination="mindpalace")
```

### Persist File
```
persist.file("/path/to/output.md")
persist.file(artifact, tags=["code", "council"])
```

### Persist Conversation
```
persist.conversation()  # Current chat
persist.conversation(summary=True)  # With auto-summary
```

### For Humans
```
"Save this whole conversation to memory"
"Persist the council responses with tag 'skill-cc-evaluation'"
"Remember this file for later"
```

---

## Platform Routing

| Platform | Destination Options |
|----------|---------------------|
| Claude.ai | Mind Palace API, download prompt |
| Claude Code | ~/.mindpalace/, local files, API |
| API | Webhook callback, return in response |
| MCP | persist_to_memory tool |

---

## Configuration

```yaml
# ~/.skillcc/persist.yaml

# CANONICAL (the vision):
default_destination: skill.cc/memory        # ← THE GOAL
default_tags: [auto-persisted]

# CURRENT IMPLEMENTATION:
# Until skill.cc infrastructure deployed:
current_endpoint: https://your-tunnel.loca.lt/api/ingest  # ← TEMPORARY

auto_persist:
  council_outputs: true
  code_artifacts: false
```

> **Note:** `skill.cc/memory` is the canonical persistence endpoint.
> Current tunnel is a bridge until infrastructure is live.

---

## Integration with Memory

```
persist → skill.cc/memory → Mind Palace
                ↓
        Retrievable via:
        - memory.search(query)
        - conversation_search tool
        - Mind Palace API
```

---

## Workflow Example

```
1. Mount skill.cc/council
2. council.convene("Evaluate skill.cc architecture")
3. [8 responses generated]
4. persist(council_output, tags=["skill-cc", "council", "strategy"])
5. Later: memory.search("skill.cc council") → retrieves this
```

---

## Related Skills

- `skill.cc/memory` — The persistence backend
- `skill.cc/council` — Generates outputs worth persisting
- `skill.cc/craft` — Build custom persist destinations

---

*Compute is valuable. Capture it.*

---

*Parent: skill.cc*