| name | dev-server-sandbox |
| description | Run multiple isolated xum dev-server instances (temp XUM_ROOT + free ports) |
dev-server sandbox instances
make dev-server starts the xum backend server, which uses a lockfile at:
<XUM_ROOT>/server.lock (defaults to ~/.xum-dev/server.lock in development)
This means you can only run one dev server per xum root directory.
This skill documents the repo workflow for starting multiple dev-server instances in parallel (including from different git worktrees) by giving each instance its own temporary XUM_ROOT.
Quick start
make dev-server-sandbox
What it does
- Creates a fresh temporary
XUM_ROOT directory
- Copies these files into the sandbox if present (unless disabled by flags):
providers.jsonc (provider config)
config.json (project list)
- Each file is seeded independently from the first root that has it
(
$XUM_ROOT, then leftover $MUX_ROOT, then ~/.xum[-dev], then
leftover ~/.mux[-dev] / ~/.cmux), so a root with only config.json
doesn't drop provider config
- Provider credential env vars are stripped from the server's env when they
could silently override or mismatch the intended setup: all of them with
--clean-providers (including Bedrock's AWS_REGION and
AWS_BEARER_TOKEN_BEDROCK; shared AWS credentials like AWS_PROFILE are
kept); otherwise only env vars that would shadow a seeded
entry that has an but no explicit
(API key env vars are always kept so env-key fallback still works)