22-spike Spike 7 ADD-3 Risk register update v1.2 → v1.3 CLOSED 2026-05-03
Outcome
outcome-VALIDATED on strict reading per refined-prompt v3.8 (Candidate 6 PLAN-FILE PRE-RESOLUTION LOCKED). Kill-condition NOT-MET: 0 risks have status ratings that contradict current spike evidence (kill-condition was ≥3 contradictions; 33 NO CHANGE rescore = 0 contradictions).
Conservative interpretation per T-file honesty-caveat 3: R-024 Anthropic pricing + R-029 Catala faith-tradition flagged as near-MITIGATE candidates but kept at current tier because underlying competitive / market risks persist regardless of TT-internal architectural framework changes. Spike evidence MITIGATES INTERNAL build-tractability (e.g., S2.10 Cedar 4.10 + cvc5 1.2.1 production-ready; S5 + S7 SSSOM at production scale 73 TSVs / 109K rows in Mondo) but does NOT MITIGATE EXTERNAL pricing / vendor / spec-evolution risks.
What was rescored
33 R-001..R-033 scored against current spike evidence (12 closed pre-Spike-7 + Spike 6 T-NEW tensions sweep + 6 cumulative-state docs):
- 0 MITIGATE (no risk lower than rated)
- 0 PROMOTE (no risk higher than rated)
- 33 NO CHANGE (current tier preserved)
Tier-shift summary: zero shifts on existing R-001..R-033. Working artefact at /tmp/spike-7/r-001-033-rescore.md.
5 R-NEW promotions
| R-N | Tier | Topic | Source | Evidence anchors |
|---|---|---|---|---|
| R-034 | MATERIAL | Hindu-pillar cultural-alignment | Spike 6 T-NEW-S6-12 | T49 cultural-alignment §top-3 ambiguities; £80-150K/yr per pillar run-rate; aligns with T-NEW-S6-01 → A-143 candidate at Q-015 lock-time |
| R-035 | MATERIAL | SSSOM 1.1 spec-evolution forward-incompatibility | Spike 6 T-NEW-S6-14 | T10 D1 + Spike 4 ζ-Q5 caveat #8; ~£500/~1d down-rev cost if curation_rule retired; aligns with A-147 candidate |
| R-036 | WATCH | G-IRI-VERIFY allow-list lifecycle | Spike 6 T-NEW-S6-03 | T-spike-rt218 caveats #4+#5; 4 BFO 2020 PURL flakies (BFO_0000056/057/121/178); aligns with A-145 candidate at A-130 codification |
| R-037 | WATCH | Multi-jurisdiction-customer Will-update + trigger fan-out | Spike 21 Q-016 caveat #6 + Spike 22 Q-017 caveat #6 + T60 #80 | UK + EU + US testator fan-out Year-2+; aligns with A-146 candidate |
| R-038 | WATCH | Civil-registry coverage gap | Spike 22 Q-017 evidence | GRO partner-API status TBD; Phase-1 manual self-report covers ~80%+ by IHT-statutory-weight via S8 6/6 HMRC vectors but only 9/14 = ~65% by trigger count |
3 from Spike 6 pre-identification (R-034 / R-035 / R-036) + 2 surfaced from Spike 21+22 evidence (R-037 / R-038).
Working artefact at /tmp/spike-7/r-new-candidates.md (full risk-register-format entries).
Tier net delta + aggregate
- Tier delta: +2 MATERIAL +3 WATCH = +5 total
- Aggregate count: 33 → 38 (CRITICAL 4 / MATERIAL 15 / WATCH 19)
- Frontmatter stale-count corrected: TOTAL 30 → 38 (mismatched with body-row-count of 33 pre-Spike-7; subagent surfaced this as side-finding)
Acquirer-DD impact
§3 risk-narrative + §4 architectural-debt closure gain 5 R-NEW rows demonstrating active architectural-debt-tracking discipline. Acquirer expects this signal at sale-time — risk register that grows + rescores against evidence (rather than static-stale) is positive due-diligence signal.
Reconsideration triggers preserved (T-file §7)
- R-034 Hindu-pillar — 5y partner-firm cultural-alignment evidence accumulates (Hindu / Sikh / Jewish / Islamic / Buddhist 5 pillars); evidence success → MITIGATE candidate
- R-035 SSSOM 1.1 — final spec publication retires curation_rule OR canonicalises it; canonical → MITIGATE; retired → trigger A-147 down-rev
- R-036 G-IRI-VERIFY — BFO 2020 PURL fix upstream (HEAD/GET 200 on flakies) → CLOSE candidate; allow-list grows >5% → PROMOTE to MATERIAL
- R-037 multi-jurisdiction — 5-firm Phase-1 pilot multi-jurisdiction case-rate >5% of estate-volume → PROMOTE to MATERIAL + Phase-1.5 conflict-of-laws engine pull-forward
- R-038 civil-registry — GRO partner-API publishes (not just public lookup) → MITIGATE; partner-firm rejects on civil-registry-detection grounds → PROMOTE
Plan-defect detection (v3.8-LOCKED Candidate 6 PRE-RESOLUTION)
Detected in plan v1.24:
- §4 Task 7 line 754 says “v1.1 → v1.2”
- §1.3 line 184 says “v1.1 → v1.2”
- §0 lastmod prose line 6 says “v1.1 → v1.2”
Actual current head is v1.2 (parallel-session bump independent of original plan target). Dispatch correctly redirected to v1.3. v3.8 Candidate 6 PRE-RESOLUTION defect class confirmed; recommend plan v1.25 → v1.26 patch correcting v1.1 → v1.2 → v1.3 references at REFERENCE session convenience.
Detection-rate update: 6 detected of 13 spike closures (Spikes 2 + 3 + 4 + 7 + 8 + 9 = 6; Spikes 1 + 5 + 6 + 10 + 12 + 18 + 21 + 22 = 7 zero-defect; ~46% rate post-v3.8 lock).
Cascade artefacts
| Artefact | Path / commit |
|---|---|
| T-file v1.0 | ~/off-github/library/projects/inherit/T-spike-risk-register-update-2026-05-03.md (322 lines / 40 KB) — off-github commit 1236d07 (push verified) |
| Risk register v1.2 → v1.3 | ~/testatetech/docs-strategy/docs/superpowers/specs/inherit-v2-risk-register.md (612 → 687 lines; 33 → 38 R-numbered rows) |
| /tmp/spike-7/r-001-033-rescore.md | 33 rescore rows (working artefact) |
| /tmp/spike-7/r-new-candidates.md | 5 R-NEW full entries (working artefact) |
| arch-state row | v3.39 → v3.40 §13 row Q-and-A-S7 + Changelog v3.40 — docs-strategy commit 0c11704 (push verified) |
| Plan §0 checkbox | Spike 7 ticked + plan v1.24 → v1.25 + CHANGELOG row — same docs-strategy commit |
| INDEX.md | NO bump (architectural-debt closure NOT forthcoming-Q-formulation) |
| docs-personal | NO commit (no richard-task disposition; A-N amendments associated with R-NEW land at relevant Q-lock times not at Spike 7) |
| Memory file | this file |
| MEMORY.md | new entry prepended above S6 |
| Active-work-log | Spike 7 CLOSED entry |
Methodological observations
- 14th spike of 22-spike suite to close (after Spikes 1+12+18+22+2+21+3+8+9+10+4+5+6)
- SECOND architectural-debt closure spike (Group 3 Spike 7 of Batch C)
- Group 3 architectural-debt closure BOTH Spike 6 + 7 closed COMPLETE = Batch C 2/2
- Batch C SPIKE-RUNNER session 8 closes — both spikes closed inline same session
- Wall-clock total Batch C: ~17:33 → 18:30 (~1h cascade including Spike 6 closure at 18:00 + Spike 7 authoring + closure 18:00-18:30)
- Pin-drift hook —fix encountered known nested-paren bug AGAIN at long-form arch-state lastmod prose; manually rewrote pins to canonical clean form
"vX.Y (YYYY-MM-DDThh:mm)"per parent CLAUDE.md §6 — same bug-class as Spike 6 closure - Spike 6 substrate (T-NEW-S6-12 + S6-14 + S6-03) directly transferred as R-034 + R-035 + R-036 — STRICT-ORDER value confirmed (Spike 6 → Spike 7 inline closure justified)
- Plan-defect detection rate ~46% post-v3.8 lock continues to validate Candidate 6 PRE-RESOLUTION discipline
- NEXT: Suite progress 14 of 22 spikes closed (64%); remaining unrun spikes 11 + 13 + 14-17 + 19 + 20 = 8 spikes across Groups 4 (Phase-1 Sprint S1/S2 pre-deliveries; Spike 11 utility-tree + Spike 13 pin-drift hook test), 5 (IK/IAS substrate confrontation; Spike 14 L4M + Spike 15 Anthropic Agent SDK + Spike 16 MS Agent Framework + Spike 17 Anthropic Skills publication), 6 (Deferred-Q pre-validation; Spike 19 Q-018 InheritKitBench + Spike 20 Q-015 t-box layering R2)