feat(FN-1491): add centralized agent health status utility

- Create agentHealth.tsx utility with consistent health status logic
- Update AgentDetailView, AgentListModal, and AgentsView to use centralized utility
- Add comprehensive tests for health status edge cases
- Update agents.md documentation with health status reference
- Add changeset for @gsxdsm/fusion patch release

.changeset/fix-agent-health-status-unresponsive.md

@@ -0,0 +1,32 @@
+---
+"@gsxdsm/fusion": patch
+---
+
+Fix false "Unresponsive" agent health status in dashboard views
+
+## What Changed
+
+Agent health status in the dashboard now correctly handles several edge cases that previously caused false "Unresponsive" labels:
+
+### Bug Fixes
+
+1. **Agents with monitoring disabled** — Agents with `runtimeConfig.enabled === false` now display "Disabled" instead of being falsely labeled as "Unresponsive"
+
+2. **Inconsistent timeout handling** — `AgentListModal` previously used a hardcoded 60-second timeout, while other views respected per-agent `runtimeConfig.heartbeatTimeoutMs` settings. All views now use the same centralized health evaluation logic
+
+3. **Health badges now stay current** — Dashboard views poll for fresh agent data every 30 seconds to keep health status accurate while views are open
+
+### Technical Details
+
+- New centralized `getAgentHealthStatus()` utility in `packages/dashboard/app/utils/agentHealth.ts`
+- All dashboard surfaces (AgentsView, AgentListModal, AgentDetailView) now use the shared utility
+- Per-agent timeout overrides (`runtimeConfig.heartbeatTimeoutMs`) are properly respected
+- Proper timer cleanup on unmount prevents memory leaks
+
+## How It Works
+
+Health labels are determined in priority order:
+- **Terminated/Error/Paused/Running** — based on agent state
+- **Disabled** — when heartbeat monitoring is disabled
+- **Starting.../Idle** — when no heartbeat data exists
+- **Healthy/Unresponsive** — based on elapsed time since last heartbeat vs configured timeout