feat(modules): HealthMonitor — scheduled probes with pluggable sinks (CORE-006)

[typelicious]

Adds a generic, pure-config HealthMonitor module to the OSS template. Org layovers declare endpoints + sinks in YAML; no per-org handler code.

Why

Some layovers (e.g. capacium-ops .github/workflows/health-monitor.yml) appended health-check lines to a tracked file inside the source repo and pushed them back every 3 hours, polluting main with auto-commits. This module provides the same observability via proper sinks — stdout, file (ephemeral CI path, not the repo), webhook, or a single labeled GitHub Issue on failure — without ever writing to the source repo.

What

• src/ops_engine/config_loader.py — HealthCheck, HealthSink, HealthMonitorConfig (Pydantic).
• src/ops_engine/modules/health_monitor.py — HealthMonitor.run(config) + CLI (python -m ops_engine.modules.health_monitor --config <yml>). Default User-Agent so CF-fronted endpoints don't 403 python-urllib. Exit 1 on failure.
• README: new entry in the modules table.

Org-layover usage

MyOrg:
  health_monitor:
    enabled: true
    checks:
      - { name: api,  url: https://api.example.com/health, expect_status: 200 }
      - { name: web,  url: https://example.com }
    sinks:
      - { type: stdout }
      - { type: github_issue, issue_label: health-alert, only_on_failure: true }

Tested

Live smoke against docs.capacium.xyz + api.capacium.xyz/health (both CF-fronted, both 200 with the UA fix) + a synthetic 404 path (correctly flagged failed).

Follow-up

capacium-ops will switch its workflow to call this module and drop the auto-commit step.