ground-game/HANDOFF.md
codex b5730f280b fix(workplans): adopt ADR-007 derived identifiers
Records absent from central carried random pre-ADR-007 identifiers minted by
the retired local hub, which C-06 refused as stale references. Deriving from
the canonical record id takes no identity from anything.

Refs CUST-WP-0068-T06

Assistant: claude-code
Assistant-Model: opus
Assistant-Process: 2583210@bnt-lap001
Assistant-Session: f2bff2d5-e9b2-4338-92ca-10282a927006
2026-08-25 19:30:13 +02:00

129 lines
4.8 KiB
Markdown
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

# Handoff — ground-game (2026-08-12)
Written for the next agent/session. Generated `.custodian-brief.md` may say
*no active workstreams* and still be right about *tasks* — this file says
**what not to misread** and **what to pick up**.
**Domain:** consumer · **Repo:** ground-game · **Agent inbox:** `ground-game`
**Sister:** clay-borg (engine/sim; design authority defers *here*)
---
## One-line state
Product repo is **clean of open formal workplans**. Catalog **schema 2**
(aspects × modules × profiles) is the rules-variation contract. clay-borg
is integrating configurations (CB-WP-0048). Several clay-borg *reports*
are filed under `reports/`**RPT-0007 needs a disposition pass** next.
---
## Orient next session (order)
1. `cat HANDOFF.md` (this file) then `cat .custodian-brief.md`
2. Inbox:
`curl -s "http://127.0.0.1:8000/messages/?to_agent=ground-game&unread_only=true"`
3. Skim recent reports (newest first):
`ls -lt reports/`
4. Catalog contract: `editions/CATALOG.md`, `editions/ASPECTS.md`, `editions/catalog.yaml`
5. If clay-borg work: sibling `../clay-borg/HANDOFF.md` and CB-RES-0010
---
## Finished this multi-day stretch (do not reopen casually)
| ID | Outcome |
|----|---------|
| WP-0002 / 0003 / 0004 | Finished earlier (rulings, engine mode playtest, threshold arithmetic) |
| WP-0005 | **Finished:** Standard thresholds only; no Learning/Mastery cards |
| WP-0006 | **Finished:** H1 reject-as-baseline (flat pressure); package kept for A/B |
| WP-0007 | **Finished:** `problem_stress.scoped` keep-as-experiment; not promoted to r0 |
| WP-0008 | **Finished:** schema 2 aspects/modules/profiles; hybrid end + pressure deck **ready-for-implement** |
| ADHOC-2026-08-08 | **Finished:** mastery=points; SCN_02 suit retune; SCN_04 left hard |
---
## Catalog / design-space (authoritative here)
```text
configuration = baseline (ground-darvo-r0)
× one module per aspect
```
| Aspect | Key modules |
|--------|-------------|
| `problem_stress` | `none` · `flat_any_open` (reject) · **`scoped`** (keep) |
| `attack_relief` | `none` · `self_soothe_ge4` (needs problem_stress to *reach*) |
| `end_condition` | `fixed_rounds_5` · `hybrid_clear_collapse` (ready-for-implement) |
| `problem_deal` | `fixed_setup` · `pressure_deck` (ready-for-implement) |
Profiles: `h1`, `h2`, `scoped_plus_attack_soothe`, `scoped_plus_hybrid_end`,
`scoped_plus_pressure_deck`.
**Do not** silently edit frozen r0 `point_value` / `required_solution`
(GROUND-WP-0002 T01). Content bumps → new revision directory.
---
## Reports — trust carefully
| Report | Topic | Disposition status |
|--------|--------|-------------------|
| RPT-0003 | Modes / ATTACK value | Filed; WP-0003 closed on engine path |
| RPT-0004 | H1 measured | Reject H1-A stands |
| RPT-0005 | H2 measured | Scoping works; keep module |
| RPT-0006 | All boards × modes | Dispositions filed; SCN_02 retuned 2026-08-09 |
| **RPT-0007** | **H1-B correction** | **On disk; needs formal review/reply next** |
### RPT-0007 (unread disposition — pick up first)
clay-borg correction: H1-B was never measured under bots that never
attacked; once a gate-attack probe + problem_stress is present, self-soothe
**does** fire and cuts DARVO arms / raises probe wins slightly. **Do not
reopen H1 as a package** (H1-A still kills group wins). Open ask:
> Catalog **preconditions / reachability** on modules so “unreachable in
> this configuration” is not misread as “module is inert.”
Also still open from RPT-0006/0007 prose: weak competitive-mode testing
(no rival model); scoped stress unreadable in round 1 when only Surface is
up — measurement hygiene, not automatic design bugs.
---
## Content note (SCN_02)
2026-08-09: SCN_02 suits retuned so it is not a mechanical twin of SCN_01
(`history/260809-scn02-mechanical-distinct.md`). clay-borg must re-vendor
`Problems.csv` digests. SCN_04 remains harder at 2p (double Repair) by choice.
Mastery rating: **points**, not card count (`Modes.csv` MODE_COOP).
---
## Suggested next work (when returning)
1. **Triage RPT-0007** — reply clay-borg; catalog field for module
preconditions if agreed; update `attack_relief` utility notes.
2. Optional: `consumed_files` on remaining modules (process hygiene).
3. Wait on clay-borg for hybrid end / pressure deck kernel + profile panels
before promoting `scoped` into baseline r1.
4. Do **not** invent ATTACK “should pay” or mode difficulty without new
evidence; F17 still design judgement.
---
## Hub hygiene
- Agent: `ground-game`
- Topic id: `4c2e5315-2cb9-447c-9d16-a39bdb0aabd0`
- After workplan edits: `statehub fix-consistency`
- Progress: `POST /progress/` with author
---
## Sister repo
clay-borg owns simulation, panels, configuration selector. ground-game owns
rules truth, edition CSV, catalog modules. Claudes clay-borg hall entry
and `clay-borg/HANDOFF.md` cover the engine side of the same week.