Suvra

Policy Management

Policy management in the dashboard is split across three tabs that share a common nav: Overview, Detail, Effective, and Rules (Rules Library).

Requires policies.read to view; policies.write to create, edit, delete, or push.

Overview (/dashboard/policy)

Lists every policy bound to the active tenant.

ColumnDescription
NamePolicy display name
TypeBadge: Global (SuvraPolicy) or Agent
DescriptionFree-text description
RulesCount of rules in the policy
AgentsNumber of agents bound to the policy, or "All" for the global
ActionsManage, Delete (Delete is disabled for the global policy)

Create policy — inline form with Name (required) and Description. Posts to /dashboard/policy/create and creates a new agent policy. The single global SuvraPolicy is initialized automatically on tenant bootstrap and cannot be duplicated.

The page also reminds operators: "SuvraPolicy applies to all agents. Agent policies add or override rules for specific agents."

Detail (/dashboard/policy/detail?policy_id=...)

Per-policy workbench.

Header — Policy name (h1), description, type badge ("Baseline" for global, "Agent Policy" otherwise).

Edit details — inline reveal form with Name + Description, posts to /dashboard/policy/{policy_id}/update.

Policy Rules — one card per bound rule showing:

  • Rule name + description
  • Category label badge
  • Effect badge (allow / deny / needs_approval)
  • Action type (code format)
  • Constraint rows (label + value pairs)
  • Remove button → POST /dashboard/policy/{policy_id}/rules/remove

The Add Rules button links to /dashboard/rules?policy_id={policy_id} so you can pick from the Rules Library.

Assigned Agents (agent policies only) — links to each agent's detail page.

Delete policy (agent policies only) — POST /dashboard/policy/{policy_id}/delete.

Effective View (/dashboard/policy/effective?agent_id=...)

Shows exactly which rules apply to a specific agent — baseline plus any agent policy combined.

Agent selection — dropdown, auto-submits on change.

Policy Stack — the layers being merged:

OrderPolicyTypeRules
1SuvraPolicyBaseline12
2agent-x policyAgent4

All Effective Rules — one card per rule after merge, evaluated in order. Each card shows:

  • Rule name + description
  • Effect badge
  • Source badge (Baseline or name of the source policy)
  • Category badge
  • Action type (code)
  • Constraint rows

Use this view to sanity-check what an agent will actually hit at runtime, or to prep a simulator scenario.

Related