Standup — 2026-06-27

Daily standup for 2026-06-27. What shipped:

ADR-0028 Phase 2b shipped end-to-end. Four PRs landed in sequence: #210 tagged 7 model-specific playbooks with business_model so the profile boost actually fires for real connected sites (it was returning 0 for every entry before this); #214 added the operator self-declaration onboarding step, site_contexts.declared column, and the effective_profile() read-time merge with per-axis boost gating (declared archetype boosts regardless of billing confidence; low-confidence billing axes don’t get laundered through); #217 closed the connect-later gap — operators who declare before connecting a site no longer lose that answer, via a users.declared_pending JSONB column applied exactly once on the 0→1 site transition. 2,194 tests green across all three.

Omar merged #207 and #219 — six draft operator playbooks plus a missing business_model tag on the diagnose-churn play, and added a GitHub issue template (#209) to keep future playbook batches consistent.

#223 adds the benchmarks global-brain collection — aggregate cohort norms across MemberIntel-connected sites, keyed to the ADR-0028 profile axes so the existing boost surfaces the right benchmark per operator. De-identification is enforced in code: k-anonymity floor (drop any cohort n < 10) and a fail-loud identifier scrubber on every rendered entry. ADR-0028 is amended accordingly. Counsel sign-off is the production gate; staging is unblocked.

Notable. The Phase 2b boost was inert from day one — no global content carried business_model tags, so applicability_boost returned 0 for every real site. The coaching demo worked only with a hand-set archetype. #210 closed that gap without a code change to retrieval.

---

Window: 2026-06-26T13:00

→ 2026-06-27T13:37

.077Z · Sources: memberintel @ c6ed294, memberpress-intel @ 9ff87ab