feat: unified configuration schema with hierarchical global defaults #511

Problem

Settings are fragmented across too many layers: host-side .env files, git-committed config/settings.yaml, container-level hub.env variables, and stateful database rows in hub.db. This creates several pain points:

• The DB Trap: Configuring a "global" service account key across all projects originally requires direct SQL database manipulation (hub.db) because there is no clean file-based config or CLI flag to set it globally.
• Redundancy: Credentials and GCP settings must be defined repeatedly on a per-project basis, violating DRY principles.
• Fragility: Settings can be silently overwritten on deploy (see bug: gce-start-hub.sh --full overwrites settings.yaml, destroying plugin config #475), lost on restart (see Admin settings saved via web UI are lost on pod restart in Kubernetes deployments #147), or ignored when defaults aren't respected (see default_harness_config is not respected after grove is init #212).

This affects all deployment targets (GCE, Kubernetes, workstation, NAS) — not just any particular environment.

Proposal

(a) Unified Configuration Schema

Consolidate all configuration into a single, comprehensive schema-validated YAML file (e.g., settings.yaml). Provide a fully documented master template (settings.yaml.example) detailing every key (global credentials, API keys, network settings, telemetry configurations) so users never have to guess or search design docs.

(b) Hierarchical Global Defaults

Introduce a global_defaults block in the unified settings configuration:

server:
  global_defaults:
    auth:
      selected_type: vertex-ai
      service_account_key_path: /secrets/credentials/credentials.json
    gcp:
      project_id: my-project
      region: us-east5

Any project or agent provisioned without explicit auth values should automatically inherit these defaults, eliminating redundant setup.

Related Issues

• bug: gce-start-hub.sh --full overwrites settings.yaml, destroying plugin config #475 — gce-start-hub.sh --full overwrites settings.yaml
• Admin settings saved via web UI are lost on pod restart in Kubernetes deployments #147 — Admin settings lost on pod restart in Kubernetes
• default_harness_config is not respected after grove is init #212 — default_harness_config not respected after init
• feat: read/write project config as agent-accessible tool #473 — Read/write project config as agent-accessible tool