Memory question

Hi, I don't use GitHub much, this is my first question so please excuse me...

I've seen that you can drive tasks forward when the Claude Code limit has been reached.

I understand the system but unfortunately not how exactly it works, would someone be so kind and could explain it to me?

```
● claude-flow - memory_usage (MCP)(action: "store", key: "migration/agents-spawned", value: "{\"timestamp\": \"2025-01-24\", \"agents\": 10, \"status\": \"all-agents-active\", \"parallel-execution\":
true}")
⎿  {
"success": true,
"action": "store",
… +8 lines (ctrl+r to expand)

● claude-flow - swarm_status (MCP)(swarmId: "swarm_1753357856324_w3vdil2e4")
⎿  {
"success": true,
"swarmId": "swarm_1753357856324_w3vdil2e4",
… +8 lines (ctrl+r to expand)
⎿  Claude usage limit reached. Your limit will reset at 4pm (Europe/Berlin).

• /upgrade to increase your usage limit.
```

What exactly do I enter as a command in Claude Code so that these tasks are assigned. Or do I have to set something first?

I have also installed even better-sql.
```
node test-sqlite.js
✅ SQLite is working!
```
Is there anything else I need to do?

Another question is, How do I save context and load this context?

I use WSL on Windows. Claude Flow Directly in Claude Code

EDIT:
After a certain time, /mcp shows that Claude-Flow is disconnected. swarm is working (connected). Is this normal?

Start with the Claude Code entry points shown in the issue: /mcp, memory_usage, and swarm_status. Reproduce the setup under WSL, including node test-sqlite.js, and investigate the reported Claude-Flow disconnection. Done means the commands, context save/load steps, prerequisites, and disconnection behavior are documented.

Written by the indexing model from the issue text.