v0.13.0

Added

• MCP tool annotations derived from each tool's governance risk_level, so
  clients can gate approval from tools/list alone. low tools advertise
  readOnlyHint: true; every other tool advertises readOnlyHint: false and
  destructiveHint: true. The medium tier includes deletes and hard
  power-offs, so a tool is never advertised as non-destructive on the strength
  of its tier. idempotentHint only repeats an explicit declaration. The exact
  level is published in _meta under io.github.AIops-tools/riskLevel. These
  are hints for the client; audit and undo are unchanged.

Fixed

• healing_health: missing metrics read as reassuring zeros. A set whose
  healing gauge was not scraped reported healingDrives: 0 (not healing); an
  absent offline-drive metric reported drivesOffline: 0; and a scrape with no
  erasure-set metrics at all (for example a metric-name mismatch) reported
  healthy: true with nothing measured. These are now null, with
  HEALING_STATUS_UNKNOWN per set and an ERASURE_SETS_UNREAD warning, and
  healthy is null when no quorum data was read. Measured zeros still read 0.
• capacity_rca: the same missing-as-zero shape — unscraped offline-drive and
  offline-node counts read 0, and a scrape with no capacity metrics reported
  healthy: true. Both counts are now null when not measured (matching
  healing_health), and healthy is null when capacity was not read.