Prevents upgrade surprises by running version-aware regression checks before you touch production. Exposes tools to snapshot your current deployment state, query a catalog of known issues (8 OpenClaw regressions grounded in field reports), generate step-by-step upgrade guides with applicable mitigations, then verify the post-upgrade state by diffing against your pre-upgrade snapshot. Surfaces new failures separately from pre-existing issues and recovered bugs. The v1.2 provider-side detection tools fingerprint API latency and response patterns to catch silent provider-side changes like Anthropic's March 2026 reasoning effort regression. Read-only by design. You still run the upgrade commands yourself. Reach for this when you manage OpenClaw or similar AI runtime deployments and need regression intel before committing to a version jump.
claude mcp add --transport stdio temurkhan13-openclaw-upgrade-orchestrator-mcp -- uvx openclaw-upgrade-orchestrator-mcp