milestones.

the 6 milestones currently driving priority. tracked in vault/00-Milestones.md.

cosmo v2 mvp + apollo auditor queue

-34d remaining

**Mon 2026-05-11** · 4 calendar days from today.

Target Mon 2026-05-11 · 4 calendar days from today.

What ships - Cosmo V2 production frontend: weekly board polish, manage-quotas modal, lane day-count interactions, backlog group counters, right-click context menu, ply-detail panel , top P0/P1 items from vault/08-Design-specs/weekly-v2-polish-backlog.md.

Apollo Auditor Queue: Gili (developer) builds from Figma (auditor-priority-v4 + edge cases). Prototype work (cosmo-design/apollo/screens/auditor-priority-minimal.html, task-card-playground.html) feeds into the Figma; HTML prototypes themselves do not ship.
Shared task model alignment between Cosmo and Apollo (per vault/08-Design-specs/cosmo-v2-mvp.md).

Owners Ishai (design) · Omer Duskin (Cosmo eng) · Gili (Apollo eng) · Sergey (product) · Yuval Balak (Apollo eng lead).

Blockers / open questions see vault/00-Action-Items.md Active Board, anything tagged [m1].

Key specs - vault/08-Design-specs/cosmo-v2-mvp.md , overall PRD

vault/08-Design-specs/weekly-v2-polish-backlog.md , V2 polish punch list
vault/08-Design-specs/2026-04-28-apollo-priority-prd.md , Apollo priority spec
vault/08-Design-specs/apollo-behavior-spec.md , Apollo behavior

cosmo remaining pages

-20d remaining

**TBD** · proposed 2026-05-25 (~2 weeks post-MVP). Confirm.

Target TBD · proposed 2026-05-25 (~2 weeks post-MVP). Confirm.

What ships (refined 2026-05-06 + 2026-05-07 brain dump) - PG Operational page (replaces Delivery, AB-027, AB-059). New top-level surface where PGs do their job. Role-routing: TLs land on Weekly, PGs land on PG Operational. Provisional silo (validate in PG interview). Absorbs Ideation's Request Ideas + Add Prompt actions + Game Page PG-context.

Settings panel (large by design, AB-028). Sub-sections: Allocations (4 matrices, including Game↔TL per P64) · Entities CRUD · Games CRUD · Builds activity log · Orphans · Schedules · Imports · Feedback archive (TBD home) · Admin preferences · Agents (future slot).
DS list/table primitive (AB-058). Lists across V1 look inconsistent. Author canonical primitive in production DS.
Auditor surface clarification , decide whether Apollo Queue absorbs Cosmo's auditor-facing JTBD or a new Cosmo surface is needed.
Nav restructure (drop Status, Dashboard, Ideation, Auditors, Builds, Feedback, McSett-as-page from Cosmo; observability moves to M3).

Pre-work - Interview Elad Verbel (PG) Sunday 2026-05-10 (AB-026) , Sergey-recommended deep knowledge source. Walkthrough of PG day-to-day before AB-027/AB-059 design. Layer INQ-001 (effort estimation, PG view) + INQ-007 (PG day-to-day pain) + role-silo question.

Sunday 2026-05-10 plan presentation to Sergey (AB-042, SUB-003). 3-bucket sitemap + 1-2 PG Operational layout directions.

Subject vault/11-Subjects/SUB-003-cosmo-ia-restructure.md.

cosmo observability consolidation

-33d remaining

**TBD** · proposed start immediately post-V2 launch (2026-05-12 onward). Ramp parallel to M2.

Target TBD · proposed start immediately post-V2 launch (2026-05-12 onward). Ramp parallel to M2.

Why now per Ishai brain dump 2026-05-07, observability is "the next thing we need to start working on" after V2 ships. Promoted from former M6 slot.

What ships - Consolidate Cosmo's currently-scattered observability surfaces into one observability area per user role.

Pages folding in: Status, Dashboard, Auditors, Revisions (analytics half), Scorecard (kept-as-is, design-pattern source), Builds (history half), Feedback (analytics half), Ideation (numbers view), McSett.
Per-role landing surfaces: TL · PG · TPM · CTO · Auditor · Head-of-PG.
Pre-work: JTBD inventory landed (vault/03-Cosmo-pages/02-observability-jtbd-2026-05-07.md) + outreach interviews to fill gaps (12 questions captured).

Spec vault/08-Design-specs/cosmo-observability.md (skeleton 2026-05-07, AB-057).

Top findings already captured - Auditors page Monday-morning JTBD (Tatiana, P63) = highest-leverage unmet observability JTBD.

6 of 8 obs pages share un-filterable / all-time-not-current / ID-mismatched DNA. P29 (revision count 4 ways) = trust-blocker.
Scorecard cleanest of the 8; use as design-pattern source.
McSett = shipping autonomous fix-bot, not placeholder. Page = observability of an autonomous co-worker.
Salvageable primitive: McSettActivityTimeline.tsx (generic per-entity audit log).

Out of scope - Operational surfaces (Weekly, PG Operational, Settings) , covered by M2 / SUB-003.

Settings:Builds-activity-log goes to Settings, not observability.
User-facing feedback engine (M1 / vault/04-Concepts/feedback-engine.md).

client platform

-34d remaining

**TBD** · activity ramps up week of 2026-05-11. Ishai's involvement target: from step 1, not passive.

Target TBD · activity ramps up week of 2026-05-11. Ishai's involvement target: from step 1, not passive.

What ships (per Sergey daily call 2026-05-06) - Brand-new customer-facing system. Customers consuming iterations leave feedback there.

Net-new project, separate from Cosmo. Sergey leading.
Cosmo will integrate / surface relevant data into it but the platform itself is new.

Ishai's stake (AB-029) - Active design involvement from kickoff, not waiting for Sergey to hand over visuals later.

Surface design + UX touchpoints before the platform's IA is locked.
Need a kickoff conversation with Sergey to scope where design fits.

Blocker Sergey hasn't named the design hand-off model yet. Ishai must push for involvement.

ux showcase site v1

-13d remaining

**TBD** · proposed 2026-06-01 (~3 weeks post-MVP). Could-do, not blocking.

Target TBD · proposed 2026-06-01 (~3 weeks post-MVP). Could-do, not blocking.

What ships - Internal Sett-team site at .pages.dev (Cloudflare Pages, matches cosmo-design.pages.dev setup).

v1 sections: Glossary · Design System showcase · Prototype Gallery (tile view with last-updated metadata, change summary, "open in new tab") · JTBD library · Daily Digest archive · Milestone board · Ventures (auto-rendered SUB-* flagged share: site-a).
Comments via Giscus (GitHub-backed, free).
Audience: Sett team + occasional show-and-tell; @sett.ai SSO.

Spec vault/08-Design-specs/sett-ux-showcase-site.md.

personal studio site v1

1d remaining

**TBD** · proposed 2026-06-15 (~2 weeks after Site A). Could-do, not blocking.

Target TBD · proposed 2026-06-15 (~2 weeks after Site A). Could-do, not blocking.

What ships - Personal studio site at .pages.dev (separate Cloudflare Pages project from Site A).

Audience: Ishai only. SSO restricted to ishairevach@sett.ai.
v1 sections:

- Subjects shelf (vault/11-Subjects/) , cards per SUB-NNN, sortable by last_touched, filterable by status/milestone - Inquiries shelf (vault/12-Inquiries/) , status filters, questionnaire view, data-gathered timeline - Action Board mirror , visual render of vault/00-Action-Items.md Active Board - Activity log live view , last 14 days from vault/10-Activity-Log/ - Prototypes hub , tile per prototype with cloud + local + Figma + push-status links (action-rich, not show-only) - Slash command launcher , gallery of /morning, /comms, /ingest-raw, /weekly-review, /brief, /daily-ingest, /inquiry - Mobile inbox status , vault/09-Notes/raw/ count + last commit on mobile-capture branch - Comms gap meter , pulled from personal/stakeholder-comms.md (gitignored) - Milestone burndown , bars per active milestone - Glossary , slash commands + keyboard shortcuts

v1 observability graphs (Recharts): subjects opened/resolved per week · decisions per week · time-to-first-decision · activity-by-surface · comms cadence heatmap · milestone burndown

Why distinct from M5 different audience (Ishai-only vs team), different surface (action-rich hub vs show-and-tell), different feature set (subjects shelf, slash launcher, observability , none on Site A). Stack shared (Astro + MDX + Cosmo DS), deploys separate.

Concept doc vault/04-Concepts/personal-studio-site.md.

First subjects bootstrapped SUB-001 (studio site itself), SUB-002 (output-to-files rule).