test(dashboard): fix flaky test + add coverage for untested services

[xsovad06]

Objective

Fix the flaky test_server_status_shows_not_running test and add dedicated test files for the highest-risk untested dashboard services.

Findings

[P1] Flaky test: test_server_status_shows_not_running

• Location: tests/test_scheduler.py:423
• Issue: Assumes no SOVA server is running, fails when dashboard is active (PID detected). Environment-dependent.
• Recommendation: Mock read_pid_file() to return None, or use unique tmp PID file path.

[P3] 28 dashboard service modules lack dedicated test files

• Location: sova/dashboard/services/ vs tests/
• Issue: Services like agent_lifecycle (1224 LOC), agent_handoff, agent_recovery have no dedicated test files. Tested indirectly via test_dashboard.py.
• Recommendation: Prioritize test files for agent_lifecycle, agent_handoff, and agent_recovery (critical auto-handoff paths).

Files to Modify

• tests/test_scheduler.py
• tests/test_agent_lifecycle.py (new)
• tests/test_agent_handoff.py (new)
• tests/test_agent_recovery.py (new)

Estimated Effort

8 hours

Dependencies

#642 (Dashboard service splits)

---

Source: docs/HEALTH-AUDIT.md -- Health Audit 2026-08-12

Triage Assessment

Title: test(dashboard): fix flaky test + add coverage for untested services
Has description: yes
Suitability: ready
Confidence: 85%
Complexity: complex
Missing context: none
Labels: type: task, area: dashboard

Issue has structured sections indicating clear scope; ready for research.