[littlehand26]

Problem

A non-validator pipeline writer can leave startup block sync while still far behind the live chain. CometBFT has no supported runtime path back from consensus to block sync, so historical catch-up then proceeds near consensus speed.

Solution

• pin feat(blocksync): add continuous sync mode cometbft#1, based exactly on the current Cronos CometBFT commit
• expose blocksync mode auto or continuous in config.toml
• keep auto as the default
• document validator restrictions, RPC semantics, monitoring, validation, rollout, and rollback

Tests

• go mod verify
• go test ./cmd/cronosd/cmd -count=1
• make build LEDGER_ENABLED=false
• make test LEDGER_ENABLED=false
• fresh cronosd init generated blocksync mode auto

Dependency

• feat(blocksync): add continuous sync mode cometbft#1

Remaining release gates

• RocksDB image build
• local multi-node and lihe-dev soak test
• production canary by SRE