feat: advance LOOP-WP-0002/0003/0004 without activity-core blockers

Document event payload contract, draft disabled ActivityDefinitions for
CI/regulator/registry hygiene, initialize loop health snapshots and
regulator dashboard, record DEC-005 kaizen-stack promotion, and close
supplier-independent workplan tasks.

activity-definitions/hourly-ci-health-escalation.md

@@ -0,0 +1,46 @@
+---
+id: coulomb-hourly-ci-health-escalation
+name: Hourly CI Health Escalation (coulomb-loop bootstrap)
+enabled: false
+owner: coulomb-loop
+governance: coulomb_social
+status: proposed
+trigger:
+  type: cron
+  cron_expression: "40 * * * *"
+  timezone: Europe/Berlin
+  misfire_policy: skip
+context_sources:
+  - type: kaizen
+    query: discover_kaizen_projects
+    params:
+      marker: .kaizen/schedule.yml
+      roster: /home/worsch/coulomb-loop/loops/kaizen-stack/roster.yaml
+    bind_to: context.projects
+---
+
+# Hourly CI Health Escalation (bootstrap)
+
+LOOP-WP-0002 T06 — pilot-only CI degradation signal. Deferred to daily phase for
+full Gitea API integration; bootstrap uses manual probe task template.
+
+```rule
+id: flag-ci-degraded
+for_each: context.projects
+bind_as: p
+condition: 'context.p.ci_failure_streak >= 2'
+action:
+  task_template: "CI degraded on {context.p.repo} — test-maintenance review"
+  description: |
+    Pilot repo shows ≥2 CI failures in 24h (signal from state-hub or probe).
+    kaizen-agentic schedule prepare test-maintenance --target {context.p.root}
+    Run failing tests locally: make test
+    Load agents/agent-test-maintenance.md
+  target_repo: context.p.repo
+  priority: high
+  labels: ["kaizen", "quality-escalation", "test-maintenance", "coulomb-loop", "bootstrap"]
+```
+
+**Note:** `ci_failure_streak` requires a future state-hub query or shell resolver.
+Keep `enabled: false` until probe resolver ships. Daily variant: rename cron to
+`0 6 * * *` when promoting per `loops/quality-escalation/cadence.yml`.

activity-definitions/hourly-loop-health-collector.md

@@ -0,0 +1,46 @@
+---
+id: coulomb-hourly-loop-health-collector
+name: Hourly Loop Health Collector (coulomb-loop regulator)
+enabled: false
+owner: coulomb-loop
+governance: coulomb_social
+status: proposed
+trigger:
+  type: cron
+  cron_expression: "50 * * * *"
+  timezone: Europe/Berlin
+  misfire_policy: skip
+context_sources:
+  - type: state-hub
+    query: recently_on_scope_hourly
+    params:
+      range: 1h
+      active_only: true
+    bind_to: context.hub_activity
+    required: false
+  - type: kaizen
+    query: discover_kaizen_projects
+    params:
+      marker: .kaizen/metrics
+      roster: /home/worsch/coulomb-loop/loops/kaizen-stack/roster.yaml
+    bind_to: context.pilot_metrics
+---
+
+# Hourly Loop Health Collector
+
+LOOP-WP-0004 T03 — aggregates first-order loop health for regulator sessions.
+Cron `50 * * * *` runs after kaizen stack (08–10 daily) and quality sweep chain.
+
+```instruction
+id: append-health-snapshot
+description: |
+  For each loop in coulomb-loop/loops/*/cadence.yml, append one JSON line to
+  loops/<loop-id>/health.jsonl with:
+  - timestamp, phase, tasks_created (from hub), manual_rescues (bootstrap-log),
+    agent_success_rate (pilot metrics avg), false_positive_rate (estimate).
+  Regenerate loops/regulator/dashboard.md from last 24h of health.jsonl files.
+  If demotion trigger from state-machine.md fires, create urgent optimization task.
+```
+
+Keep `enabled: false` until instruction executor supports file-append actions.
+Manual collection documented in `loops/regulator/session-template.md` (T05).

activity-definitions/hourly-registry-hygiene-sweep.md

@@ -0,0 +1,47 @@
+---
+id: coulomb-hourly-registry-hygiene-sweep
+name: Hourly Registry Hygiene Sweep (coulomb-loop bootstrap)
+enabled: false
+owner: coulomb-loop
+governance: coulomb_social
+status: proposed
+trigger:
+  type: cron
+  cron_expression: "20 * * * *"
+  timezone: Europe/Berlin
+  misfire_policy: skip
+context_sources:
+  - type: state-hub
+    query: repo_sbom_status
+    params:
+      repos: pilot
+    bind_to: context.repos
+  - type: shell
+    query: reuse_surface_report_gaps
+    params:
+      roster: /home/worsch/coulomb-loop/loops/registry-hygiene/rollout.yaml
+      batch_size: 3
+    bind_to: context.gaps
+---
+
+# Hourly Registry Hygiene Sweep
+
+LOOP-WP-0003 T02 — round-robin batch from `rollout.yaml` (3 repos/hour).
+
+```rule
+id: hygiene-task-per-signal
+for_each: context.gaps
+bind_as: g
+condition: 'context.g.hygiene_signal != ""'
+action:
+  task_template: "Fleet hygiene: {context.g.signal} on {context.g.repo}"
+  description: |
+    Follow loops/registry-hygiene/session-template.md
+    kaizen-agentic schedule prepare scope-analyst --target {context.g.root}
+    reuse-surface validate
+  target_repo: context.g.repo
+  priority: medium
+  labels: ["fleet-hygiene", "coulomb-loop", "scope-analyst"]
+```
+
+`enabled: false` until `reuse_surface_report_gaps` shell resolver exists in activity-core.