09 — RP/UI Deploy Packet (Workstream H)

Patch-ready over the v4 views. Counts come from PG views — no Nuxt math, verified badge only from v_process_discovery_verified_candidates_v3. Operator-gated deploy (Nuxt rebuild not performed here).

Routes (read-only, over v4 views)

route	source view
GET /process-discovery/runner-status	v_process_discovery_runner_status
GET /process-discovery/fixtures	process_run_observation where evidence_ref->>'fixture_ref' IS NOT NULL
GET /process-discovery/no-execute-tests	process_run_observation where source_system='dryrun_prepare_runner'
GET /process-discovery/dry-run-readiness	v_process_discovery_dryrun_readiness
GET /process-discovery/candidate/:code/runtime	v_process_discovery_candidate_status_v4 filtered
GET /process-discovery/candidate/:code/observation-plan	run+component obs for the candidate
GET /process-discovery/candidate/:code/components	process_component_observation for the candidate

Badges (derived, not computed in Nuxt)

• runner_missing ← dryrun_capability='requires_runner'
• runner_contract_missing ← dryrun_capability='requires_agent_api_contract'
• fixture_ready ← fixture_ready=true
• no_execute_tested ← no_execute_tested=true
• dry_run_ready ← dryrun_readiness_state='dry_run_runner_available'
• dry_run_observed ← dry_cnt>0
• verified ← row present in verified_candidates_v3 (ONLY source). dot:kg must never show verified.

Anti-fake rules

1. Verified badge sourced solely from the v3 verified view.
2. SIMULATED/fixture/no-execute render as their own neutral badges, never green-verified.
3. job:cut shows dry_run_runner_available (capability) — not "dry-run done".
4. No client-side aggregation of counts.

Deploy = operator Nuxt rebuild (server API handlers + badge components). Backend (views) is live and stable.