fix: get_indexing_status lastUpdated remains stale after incremental sync

[linze0721]

Summary

Fixes #395: incremental sync (handleSyncIndex → reindexByChange) now updates persisted MCP snapshot metadata so get_indexing_status reflects recent sync activity, not only the last full index.

Root cause

SyncManager.handleSyncIndex() completed vector/Merkle updates but never called SnapshotManager to bump freshness or save ~/.context snapshot JSON. get_indexing_status read indexedInfo.lastUpdated, which was only set on full index / recovery paths.

Changes

Area	Change
config.ts	CodebaseInfoIndexed: optional lastFullIndexAt, lastIncrementalSyncAt, lastSyncStats
snapshot.ts	recordIncrementalSyncSuccess() updates lastUpdated, incremental fields, optional file/chunk counts; setCodebaseIndexed() sets lastFullIndexAt
sync.ts	After successful reindexByChange per codebase, record stats and saveCodebaseSnapshot()
handlers.ts	Status text: Last full index, Last incremental sync (with stats or “no file changes”), Index freshness (lastUpdated)
Tests	snapshot.incremental-sync.test.ts, sync.incremental-metadata.test.ts; extended handlers.get-indexing-status.test.ts

Verification

pnpm install --frozen-lockfile && pnpm --filter @zilliz/claude-context-mcp... run build && pnpm --filter @zilliz/claude-context-mcp test

7/7 tests passed.

Commit

• SHA: 563063b
• Branch: berserk/e3b41593-0b22-40ec-a8a8-9181c1d52479 (pushed to origin)

Client impact

Agents can distinguish: stale full-index-only timestamp vs successful incremental sync (including zero-change runs) without forcing index_codebase force=true.

Commit

563063b

Branch

berserk/e3b41593-0b22-40ec-a8a8-9181c1d52479

Fixes #395