diff --git a/workplans/GROUND-WP-0007-h2-scoped-problem-stress.md b/workplans/GROUND-WP-0007-h2-scoped-problem-stress.md
new file mode 100644
index 0000000..4f33a3b
--- /dev/null
+++ b/workplans/GROUND-WP-0007-h2-scoped-problem-stress.md
@@ -0,0 +1,108 @@
+---
+id: GROUND-WP-0007
+type: workplan
+title: "H2 experiment: scoped problem stress (personal / bond / global)"
+domain: consumer
+repo: ground-game
+status: active
+owner: bernd
+topic_slug: whynot
+created: "2026-08-08"
+updated: "2026-08-08"
+# measurement: GROUND-RPT-0005
+origin: residual
+origin_ref: GROUND-WP-0006
+state_hub_workstream_id: "6a72a7c3-b408-4f3a-9f15-bf204e52653d"
+---
+
+# H2 — route problem stress by scope; measure bond-network incentive
+
+**Parent:** [GROUND-WP-0006](GROUND-WP-0006-h1-problem-stress-experiment.md)
+(H1 reject-as-baseline). Design:
+[`history/260808-h2-scoped-problem-stress.md`](../history/260808-h2-scoped-problem-stress.md).
+
+**Package:** [`editions/experiments/h2-scoped-problem-stress/`](../editions/experiments/h2-scoped-problem-stress/)  
+**Catalog id:** `h2-scoped-problem-stress`  
+**Base:** `ground-darvo-r0` (do **not** stack H1-A/H1-B).
+
+Bond-scope intent: bonded seats share End stress on that card, so the
+network is jointly incentivised to SOLVE it (any seat may still SOLVE).
+
+## Task: freeze H2 package and catalog entry
+
+```task
+id: GROUND-WP-0007-T01
+status: done
+priority: high
+state_hub_task_id: "8e492c40-f0d2-4b26-990d-c01e87f861e5"
+```
+
+Land experiment directory, `rules_delta.yaml`, catalog row, design note.
+Baseline r0 and H1 package untouched (H1 stays reject-as-baseline control).
+
+**Done 2026-08-08.**
+
+## Task: request clay-borg implementation and panel
+
+```task
+id: GROUND-WP-0007-T02
+status: done
+priority: high
+state_hub_task_id: "4b33a9ac-5b97-488f-bf0e-3a352053d1b0"
+```
+
+Message clay-borg: select `h2-scoped-problem-stress`; implement H2-SCOPE,
+H2-OWN, H2-A, H2-SOLVE; **exclude** H1 deltas. Re-run attack-value /
+regulation-style panel vs `ground-darvo-r0` and optionally `h1-problem-stress`.
+Ask for: group win rates, stress variance (max−min), DARVO arm rates by
+policy, and if cheap **SOLVE rate / time-open on bond-scoped vs personal**
+cards under Bonded tables.
+
+**Done 2026-08-08.** Hub message to clay-borg-custodian requesting panel.
+
+## Task: human table smoke (optional)
+
+```task
+id: GROUND-WP-0007-T03
+status: todo
+priority: medium
+state_hub_task_id: "116be9c8-d0e0-4da7-b9e4-15c531fff629"
+```
+
+2p (no bond card) and 3–4p (bond card in play). Note: does Bond feel like
+shared liability worth clearing; does personal stress create asymmetric
+DARVO without whole-table freeze.
+
+## Task: record measurement and catalog utility
+
+```task
+id: GROUND-WP-0007-T04
+status: done
+priority: high
+state_hub_task_id: "b26e457e-fa2b-466c-8fb4-3b2cc4ca86db"
+```
+
+Update catalog `status` / `utility_estimate` / `decision` from results.
+File report under `reports/` if clay-borg sends one.
+
+**Done 2026-08-08.** [GROUND-RPT-0005](../reports/260808-clay-borg-h2-measured.md);
+catalog `measured` / `keep-as-experiment` (not promote yet — deal & pacing
+open questions).
+
+## Task: decide promote / reject / H3
+
+```task
+id: GROUND-WP-0007-T05
+status: wait
+priority: medium
+state_hub_task_id: "46118988-9421-40ba-812b-46f8b15f8209"
+```
+
+Maintainer decision. Options: promote scopes into r1 design; reject;
+tune deal mix (difficulty dial); add ATTACK self-soothe only if seats
+reach ≥4 and projection is still dead.
+
+**2026-08-08 interim:** Keep H2 stress routing. Do **not** promote until
+deal influx / end conditions / sequence salience are decided (see
+session design notes). Bond SOLVE incentive still needs table or
+scope-aware bots.