Add npm run alias:{tribes,sender,kobold,all} to run minimal coverage
groups of scenarios sequentially, failing fast on the first failure.
Scenarios in a group share one SUPERTEST_RUN_ID so artifacts land
together; all is the deduplicated union of the named groups.
tribes omits single-node-plugin-rollout-rollback (its coverage is a
subset of the cluster rollout scenarios); its rollback executor-status
check is folded into cluster-plugin-rollout-sync-split-brain via a shared
assertRollbackExecutorAccepted helper.
Extract the duplicated scenario plumbing into src/scenarios/shared.ts:
init-primary preamble, rollout start, cluster control convergence,
supertest plugin schema/state helpers, and sleep/shellQuote/errorToMessage.
Add manual provider configuration and node-create wiring for importing an existing host into supertest. Document the manual-node-init scenario and cover config plus CLI argument generation in tests.
Add a two-node scenario that rolls out the Kobold plugin, creates public and private datasets, and verifies public dataset sync while private data remains local.
Add a live supertest scenario that rolls out the supertest plugin on the primary node, adds join nodes afterward, and verifies the new nodes install with the active plugin-enabled system target.
Rename the clustered Sender fanout scenario to cover the edge reboot recovery path and wire the new command through the CLI and npm scripts.
Use provider default instances from runtime config, allow cleanup diagnostics to tolerate expected timeouts, and document the updated scenario and live-debug helper guidance.
Validated with SUPERTEST_CERT_MODE=self-signed SUPERTEST_WAIT_ON_FAILURE=1 npm run scenario:cluster-sender-fanout-reboot and npm run typecheck.
Add the three-node Sender fanout scenario, wire it into the CLI and npm scripts, and document the scenario and substitute preflight helper. Capture remote logs before teardown so failed live runs keep enough context for later analysis.
self