Vibeship Spawner UI — Design skill for Claude Code

Spawner UI

Spawner UI is the execution plane and local dashboard for the Spark stack.

In the current supported starter architecture:

• spark-telegram-bot owns Telegram ingress
• spark-intelligence-builder is the Spark runtime core
• spawner-ui runs mission execution and the local visual control surface

Spawner UI does not own the Telegram bot token and does not receive Telegram webhooks directly.

Where It Fits

flowchart LR
  Telegram["spark-telegram-bot"] --> Run["/run goal"]
  Run --> API["Spawner UI APIs"]
  API --> Mission["Mission builder and executor"]
  Mission --> Canvas["Canvas / project workspace"]
  Mission --> Events["Mission events"]
  Events --> Telegram

What It Does

• provides the mission-building and mission-control UI
• exposes the local APIs used by the Telegram gateway for /run, /board, and /mission
• runs multi-step execution flows behind the gateway and Builder
• receives mission lifecycle callbacks from Spawner to Telegram through the local relay URL configured by Spark CLI
• loads PRD/project plans into the visual canvas
• coordinates configured LLM/provider runtimes through the mission orchestration layer

Current Role In The Spark Stack

Telegram
  -> spark-telegram-bot
  -> spark-intelligence-builder
  -> spawner-ui when execution is needed

Spawner UI is the execution backend in that shape, not a competing ingress surface.

Spark CLI starter setup writes:

• MISSION_CONTROL_WEBHOOK_URLS pointing at the Telegram relay
• TELEGRAM_RELAY_SECRET shared with spark-telegram-bot
• non-secret LLM provider metadata such as provider, model, and base URL

Do not put Telegram bot tokens or cloud LLM API keys in Spawner UI env unless a specific provider integration explicitly requires them.

First Run

Most users should let Spark CLI install and wire this module:

spark setup
spark start spawner-ui
spark status

Manual local development:

git clone https://github.c