v0.7.0 release-gate final — 2026-05-22 (7,321 / 0 / 0)

Headline

The numbers.

7,321passed

0failed

0ignored

269binaries

22issues closed

2substrate fixes

Test invocation: cargo test --release --no-default-features --features sal,sal-postgres,sqlite-bundled -- --include-ignored --test-threads=1. The --include-ignored flag is load-bearing: it promotes the 30 previously-gated live_* rows into the default run, which is how the #1120 pgvector substrate defect surfaced.

In-campaign fixes

22 GitHub issues closed with retest + re-check evidence.

Two QC passes (agents a92308816df776eb7 for #1120–#1129 and a561ae68f0605cb1e for #1130–#1141) verified the batch against the C1–C8 orchestrator-safeguard set.

Engineering writeups

Per-track results.

README.mdCampaign index, hard rules, namespace convention, provenance, full 22-issue table→ track-a-build-install-results.mdBinary SHA + symlink + container + env-var topology pinned for downstream tracks→ track-b-a2a-results.mdA2A 8-scenario re-verification post-22-fix; cross-track audit trail→ track-c-postgres-age-results.mdPostgres + Apache AGE full regression; v15→v49 migration ladder; cross-store parity scorecard→

Three audiences

Same campaign, three abstraction levels.

Every campaign produces three audience-facing pages plus the engineering writeup. Per operator directive 2026-05-18 pm.

For non-technical readersWhat ai-memory does, what we tested, what the verdict means — plain English, zero jargon→ For executives / PMs / decision-makersVerdict, risk profile, cost, v0.6.4→v0.7.0 comparison, roadmap impact, recommendation→ For SME engineers + architectsReproducibility contract, per-issue root-cause table, gate-matrix evidence, cross-store parity scorecard, future-bug prevention→

Discipline

Testing-loop addendum compliance.

No deferral. 22 of 22 issues fixed in v0.7.0; none labelled "non-blocking", "P2/P3", "vN+1 polish", or "DEFER-TO-V080".
Verify-before-claiming. Every "tests pass" claim cites cargo test invocation + result line; every commit SHA resolves via git show.
Substrate-vs-fixture honesty. 2 of 22 (#1120, #1134) are genuine substrate defects; 20 of 22 are test-fixture pin-update discipline. Both populations counted as real defects per pm-v3.
Retest + re-check discipline. Each fix verified via the same scenario that surfaced it (retest) PLUS a fresh-angle probe (re-check) before close.
Recompile + batch retest. Final full-suite run (v18) executed against the composite tip fd172f2cf to mint the 7,321 / 0 / 0 verdict.
QC pass discipline. Two independent QC passes (different agent IDs) verified the batch against the C1–C8 set before SHIP-RECOMMENDED was authored.

Provenance

Reproducibility contract.

Campaign date: 2026-05-22
Operator: justin@alpha-one.mobi
Authoring agent: Claude (Opus 4.7, 1M context)
QC pass 1: agent a92308816df776eb7 — verified #1120–#1129
QC pass 2: agent a561ae68f0605cb1e — verified #1130–#1141
Branch: release/v0.7.0-mobile-ci-1068 (tracks origin/release/v0.7.0)
HEAD: git SHA fd172f2cf
Binary SHA256: d4b60aa5b8f97470d95007f30bddb15e7e35c3855f0085c6b4f43d57f6b4ef3e
Container: ai-memory-lan-parity-pg-age on 127.0.0.1:15432 (PG16 + AGE 1.6.0 + pgvector 0.8.2)
Schema: v49 (sqlite + postgres ladders aligned)
Full-suite log: .local-runs/full-suite-final-v18-2026-05-22.log
Prime directive: memory cd8ede94-3376-4837-b570-9d975290ae08 (pm-v3: verify-before-claiming + no-operator-handoffs)
Release-gate issue: #836 — v0.7.0 RELEASE GATE

2026-05-18 — Post-Gap-7 dogfoodWire-schema + docstring drift sweep; 4 dogfood findings closed→ 2026-05-18 — Track A NHI re-run + fix batch10 GH issues closed; 85 PASS / 0 FAIL across 12 phases→ v0.7.0 release notesWhat changed since v0.6.4→

v0.7.0 release-gate final testing dossier.