How to correctly leverage Ruflo for successive commands?

[darrendavid]

I successfully spun up Ruflo with a swarm using claude-flow and much work was performed, however, subsequent commands in the same Claude Code window don't seem to be engaging a swarm or multiple agents - what are the best practices for successive commands? Do I need to quit out of the Claude code terminal and spin up a new swarm from the command line, or is there's something I should be adding in my prompt to make the most of Ruflo? It does seem like the session is aware of Ruflo, given the status bar:

 ▊ RuFlo V3.5 ● user  │  Opus 4.6 (1M context)  │  ⏱ 26m34s  │  ● 13% ctx  │  $1.44     
  ─────────────────────────────────────────────────────
  🏗️  DDD Domains    [○○○○○]  0/5    ⚡ target: 150x-12500x
  🤖 Swarm  ○ [ 0/15]  👥 0    🪝 16/16    🔴 CVE 0/0    💾 3MB    🧠   1%
  🔧 Architecture    ADRs ●0/0  │  DDD ●  0%  │  Security ●PENDING
  📊 AgentDB    Vectors ●23  │  Size 16KB  │  Tests ●0 (~0 cases)  │  MCP ●1/1