sync: mirror deployed github repo (chains + sanitized bridge), keep GitLab CI

Make this monorepo's tracked tree identical to the deployed github repo
(github.com/pezkuwichain/pezkuwi-DKS): all RC/AH/People runtimes and framework
carry the mainnet fixes (incl. U128CurrencyToVote everywhere, so the staking
election stall cannot recur), and tools/usdt-bridge matches the current,
public-safe custody tooling (real config/deploy units gitignored, secret-file
locations and server IPs stripped, sanitized *.example templates committed).

Only exception to the mirror: GitLab-native CI is preserved (.gitlab-ci.yml and
.gitlab/), to be rebuilt to match the github CI pipeline in a follow-up.

pezkuwi/node/service/src/builder/partial.rs

@@ -146,7 +146,7 @@ pub(crate) fn new_partial_basics(
 
 pub(crate) fn new_partial<ChainSelection>(
 	config: &mut Configuration,
-	Basics { task_manager, backend, client, keystore_container, telemetry }: Basics,
+	Basics { mut task_manager, backend, client, keystore_container, telemetry }: Basics,
 	select_chain: ChainSelection,
 ) -> Result<PezkuwiPartialComponents<ChainSelection>, Error>
 where
@@ -215,6 +215,24 @@ where
 			telemetry: telemetry.as_ref().map(|x| x.handle()),
 		})?;
 
+	// `import_queue`'s babe-worker essential task lives only as long as some clone of
+	// `babe_worker_handle` (its channel sender) is alive. For `new_full` an RPC-bound clone
+	// keeps it alive naturally, but chain_ops-only callers (e.g. `new_chain_ops`, used by CLI
+	// utility/benchmark commands) discard `PartialComponents::other` — which held the only
+	// other surviving clone — dropping the last sender almost immediately. That closes the
+	// worker's channel, the essential task completes, and it gets logged as "failed" (any
+	// completion is logged as failure by `spawn_essential`), even though nothing is actually
+	// broken. Anchor a clone to `task_manager` itself so it survives for as long as the caller
+	// keeps `task_manager` alive, regardless of which components they keep.
+	//
+	// `transaction_pool`'s own essential tasks (`txpool-background`, `transaction-pool-task-0`/
+	// `-1`) have the identical problem: `new_chain_ops` doesn't even include `transaction_pool`
+	// in its return tuple, so it's dropped as soon as `new_partial` returns unless something
+	// else anchors it. Chain_ops-only callers don't need a live pool (they talk to `client`
+	// directly), but a dead pool still trips the same "essential task failed" false alarm.
+	// Anchor it here too, for the same reason and by the same mechanism.
+	task_manager.keep_alive((babe_worker_handle.clone(), transaction_pool.clone()));
+
 	let justification_stream = grandpa_link.justification_stream();
 	let shared_authority_set = grandpa_link.shared_authority_set().clone();
 	let shared_voter_state = pezsc_consensus_grandpa::SharedVoterState::empty();