TT Claude Memory

project_batch_acquirer_dd_package_closed_2026_05_20

Properties1
descriptionBATCH acquirer-dd-package §3.x synthesis closure — A-231 + task-118 + 36-page Anthropic-acquirer-ready PDF + 5 audit-records subfiles synthesising 6 weeks of substrate; first acquirer-DD-grade synthesis batch; 3-way commit-time race-condition collision-resolved via worktree + rebase + renumber

5 min read

BATCH acquirer-dd-package §3.x synthesis — CLOSED 2026-05-20T22:55 BST

Per locked strategic direction project-strategic-direction-locked-build-and-sell-2026-05-20 (Rich-lock 2026-05-20: BUILD-AND-SELL TO ANTHROPIC; Lexis/TR/WK backup acquirers). First acquirer-DD-grade substrate-synthesis batch — packages 6 weeks of intensive substrate-accumulation into a single Anthropic-acquirer-ready PDF + supporting audit-records bundle.

Deliverables

  • PDF: /mnt/c/Users/Richard/Downloads/inherit-v2-acquirer-dd-package.pdf — 36 pages / 192 KB (post-pypdf metadata embed) / pandoc + WeasyPrint per CLAUDE.md §11.
  • 5 audit-records subfiles at docs-strategy/docs/superpowers/specs/2026-04-29-multi-phase-audit/audit-records/2026-05-21-acquirer-dd-package/:
    • README.md (entry point + bundle guide + done-criteria + final commit-time numbering)
    • source-inventory.md (verbatim upstream artefact list + commit SHAs for reproducibility; 13 sections)
    • acquirer-dd-package.md (PRIMARY DELIVERABLE; 9-section ~380-line structure per launch-prompt §0 #1)
    • extraction-methodology.md (synthesis methodology per-section + 10 SCFs folded inline with verification-chain table + 6-method §13.6 toolkit application audit)
    • acquirer-dd-package-coverage-matrix.md (T-1..T-20 technical + C-1..C-14 commercial + R-1..R-8 risk criteria with reading recommendations per acquirer role)
  • arch-state v4.23 → v4.24 with A-231 (newest-first per registry §15 convention)
  • ledger v1.65 → v1.66 with task-118 NEW + DONE inline + v1.66 CHANGELOG
  • launch-prompt added to repo for traceability

10 SCFs caught at synthesis time + folded inline per §13.5 rule 3

#Launch-prompt claimVerified at synthesisResolution
1”228 amendments”A-228 highest registry; 85 operative §15 body rowsBoth cited honestly — EMBEDDING-MATERIAL-GAP recurrence pending ν‴-prime
2”~15 cascade-Q files”40 locked + 4 archive subdirsGrew from 15 via Q-027..Q-044 + Q-CM-1 cascade
3”23 SKOS classifier schemes”95 unique *Scheme class names23 canonical baseline + ~72 module-specific Q-cascade additions
4”9 PROVEN-VIA-BRIDGE bodies”21 .catala_en files + ~17 .expected-output.json pairs9 BATCH υ baseline + 6+ candidates pending full re-run
5”33-risk register”43 risksGrew from 33 via 22-spike Spike 7 ADD-3 + post-suite additions
6”~35 memories”296 memory files total35 = MEMORY.md index entries; 296 = total files
7”22 A-21 CI gates”22 confirmedNo SCF
8”11-body standards-engagement plan”11 confirmedNo SCF
9”5 canonical specs”10+ canonical artefacts5 snapshot + 3 live state + 1 brand + 1 master-plan
10(implicit) Audit-records subdirs22 documented—

PDF post-generation verification per CLAUDE.md §11 PASS

  • ls -la /mnt/c/Users/Richard/Downloads/inherit-v2-acquirer-dd-package.pdf — 192 KB non-zero
  • pdftotext extracts all 9 §1-§5 + §3.1-§3.6 + §4-§5 section headers
  • pypdf metadata read-back confirms /Title + /Subject v1.0 + /Author Rich Davies + /Keywords + /Source fields embedded

3-way commit-time race-condition collision-resolved cleanly

  • Worktree-isolated execution at /tmp/batch-acquirer-dd from origin/main clean baseline per BATCH ο′ R1+R3 worktree-isolation precedent.
  • Initial commit attempt rejected by 2 parallel-session commits landing between worktree-create + push-attempt:
    • bb9b3dd BATCH q-046-substrate Q-CM-6 DisposalAuthorisationEvent — took A-229 + task-116 + v4.22 + v1.64
    • 9ea66a9 BATCH OASIS engagement evidence-pack — took A-230 + task-117 + v4.23 + v1.65
  • Rebase to fresh origin/main + renumbering to A-231 + task-118 + v4.23→v4.24 + v1.65→v1.66 applied per feedback-batch-iota-follow-up-commit-time-numbering-bump (BATCH ι follow-up A-222 precedent) + Decision Matrix #10 redirect-uncertainty-to-artifacts.

This is the first 3-way race observed in the storm cluster (BATCH ι follow-up A-222 was 2-way). The worktree + rebase + renumber pattern scaled cleanly to 3-way. Pattern-stability for §13.5 verify-before-author EXTENDS to 36/36 (10th conceptual application).

Anti-pattern check ALL 6 PASS per launch-prompt §4

  1. Modifying upstream artefacts during synthesis — NO; READ-ONLY discipline + SCFs folded into extraction-methodology.md
  2. Asserting counts without grep-verify — NO; every count cited has verification command in extraction-methodology.md §3
  3. Skipping PDF post-generation verification — NO; ls -la + pdftotext + pypdf metadata read-back ALL ran
  4. Authoring marketing-prose / sales-pitch tone — NO; §1.2 What-NOT + §2.4 What-NOT-claimed disclaim sales-pitch tone
  5. Touching code-inherit-v2 — NO; READ-ONLY on Catala + LinkML + SSSOM
  6. Hand-rolling PDF generation — NO; used pandoc + WeasyPrint per CLAUDE.md §11

Significance

This batch is the first acquirer-DD-grade synthesis batch demonstrating that substrate-quality discipline (verify-before-author 6-method §13.6 toolkit + 5-closure convergence MQ-014..MQ-018 + L7 RESEARCH-INLINE-AT-Q-LOCK-TIME) scales to acquirer-DD-grade evidence packaging as a downstream deliverable surface. The pattern: 6 weeks of intensive cascade-Q work produces, as a side-effect, an acquirer-ready substrate that can be consolidated in ~2.5h Claude time (1h inventory + 4h synthesis + 30min PDF + 1h commit-prep including 3-way race recovery).

Acquirer-DD substrate-baseline COMPLETE for Anthropic technical-DD package delivery 2026-Q2 per locked strategic direction project-strategic-direction-locked-build-and-sell-2026-05-20 (parent batch).

Process discipline upshot (for next acquirer-DD-grade synthesis batch)

  1. Worktree-isolation is the right default for parallel-session storm contexts. Even if you start before parallel sessions, their WIP in the parent working tree creates collision risk; worktree gives you clean origin/main baseline.
  2. 3-way races are now precedent-handled. The rebase + renumber pattern is proven at 3-way; future batches should not hesitate to use it.
  3. SCFs caught at synthesis time are honest framing assets. A package that says “228 amendments registry vs 85 operative body rows” is more credible than one that asserts “228 amendments” without context. Acquirer-trust is built on honest framing.
  4. The pattern-stability counter is a real signal. Each application of §13.5 verify-before-author + §13.6 6-method toolkit extends the counter; this batch took it to 36/36.

Cross-references

  • Parent strategic-direction memo: strategy/orgs/anthropic/anthropic-acquirer-strategic-direction-locked.md v1.0+ (Rich-lock 2026-05-20)
  • Launch-prompt: docs-strategy/docs/superpowers/specs/2026-04-29-multi-phase-audit/batch-acquirer-dd-package-section-3-synthesis-launch-prompt.md v1.0
  • Commit: docs-strategy 37623d0 (origin/main)
  • arch-state row: A-231 in §15 operative body at top
  • ledger row: task-118 at L196 in operative task table
  • Audit-records bundle: docs-strategy/docs/superpowers/specs/2026-04-29-multi-phase-audit/audit-records/2026-05-21-acquirer-dd-package/
  • PDF: /mnt/c/Users/Richard/Downloads/inherit-v2-acquirer-dd-package.pdf
  • Sibling closed batches in the storm cluster: project-batch-q-046-substrate-closed (A-229) + project-batch-oasis-engagement-evidence-pack-closed (A-230)

Graph View

Backlinks