chapter.guide / swarm / run 013

Os Tab Audit

2026-05-05  ·  CONCERNS  · Bookmark this page: /swarm/013-os-tab-audit.html

Assignment

Audit the entire Next Chapter OS tab structure and content — identify revenue features, distractions to sunset, missing tabs, orphaned artifacts, and cleanup recommendations.

Subset: Full roster — all 11

Roster

writerarchitecthunterquarterbacklistenerdrapermarket-analystaudit-qualitystorytellerdebrieftech-translator

What we found

The OS is a deal machine dressed as a developer dashboard. The two most important revenue tools — the caller app (where SDRs make calls) and deal management (where deals get tracked through stages) — are fully built but have no tab in the navigation. You literally cannot find them unless someone tells you the URL. Meanwhile, tabs like "Signals" (a scoring-criteria registry) and "Artifacts" (4 markdown files) occupy prime real estate. Only 27-33% of the navigation serves revenue features vs. the 50-60% standard in comparable deal platforms.

Additionally, 744 HTML pages of buyer research, call-cards, dossiers, and proposals sit in public/ directories with no discoverability path. 75 SDR call-cards — the actual prep sheets callers use before dialing — have zero inbound links from anywhere in the system.

Why this matters

Ewing is hiring interns and bringing on executives who haven't been building with him. When they open the OS for the first time, they see a manifesto about file management ("No hidden files") instead of a revenue dashboard. They see 15 tabs of equal visual weight with no story about how deals flow. They cannot find the dialer. They cannot see the deal pipeline. They don't know 744 pages of intelligence exist.

The highest-impact fix is also the simplest: add "Deals" and "Caller" to the primary navigation. Two lines of code in layout.tsx. Everything else — Dashboard transformation, tab merging, deal war rooms — can follow in tiers.

Where we agreed

All 10 agents independently reached the same top-3 findings:

1. Caller and Deals must be promoted to primary navigation (unanimous)

2. Home page must become a live dashboard, not a manifesto (unanimous)

3. Lists vs. Company Lists naming confusion must be resolved (unanimous)

This was the strongest consensus in 13 runs.

Where we disagreed

Minor naming variations only — no substantive dissent:

  • Swarm rename: "Agents" (tech-translator), "Strategy Team" (draper), keep "Swarm" (writer). Recommendation: "Agents" — clearest.
  • Tab count target: ranged from 7 (draper) to 13 (writer). Recommendation: 11 (7 primary + 4 secondary).
  • Mark/Ewing tabs: most said demote, draper said evolve to "My Desk." Recommendation: link from Dashboard, not primary nav.

What surprised us

  • The caller app (/caller) has 930 lines of production-quality frontend — login, dashboard, call pages, outcome logging, daily summaries — and zero navigation presence. It is the single largest usability gap in the system.
  • HR.com content exists at 7 different URL patterns. A single deal has 7 entry points.
  • The Admin/Dashboard page is actually the best deal cockpit in the system — it already queries v_deal_pipeline and shows stages, fees, probabilities. It just needs to be called "Deals" instead of being buried under "Admin."
  • 75 SDR call-cards (public/projects/hrcom/call-cards/) are purpose-built revenue tools with zero links from any page.
  • apps/deals/ contains 171 files of a dead Vite/React app completely superseded by the Next.js /deals/[slug] route.

What we'd do differently

  • Start by rendering the actual pages in a browser, not just reading code — the emotional/usability audit benefits from seeing what users see.
  • Query Supabase directly for pipeline state before tracing UI connections — the data model reveals more about system intent than code structure.
  • Check for existing implementations before proposing new builds (run #010's lesson, confirmed again here).

Currency events

From → ToActionMultiplierBaseScoreNotes
architect → allRevenue pipeline gap map: 4 stages, quantified connected vs disconnected3x1030First systematic mapping of Marketing→Selling→Closing→Executing chain
hunter → caller-team75 orphaned call-cards identified as highest-priority navigation gap3x1030Revenue tooling with zero discoverability
quarterback → conductor5 critical-path breaks identified in lead→close pipeline chain3x515Break points: no pipeline view, hidden caller, no deal index, silent review, dark executing
listener → draper"Three products sharing one URL" insight: OS, Sales Engine, Client Layer3x515Root cause of nav complexity
storyteller → writer"Protagonist must be the deal" narrative framework for tab ordering3x515Deal lifecycle becomes the story arc
market-analyst → ewing$200-400K/quarter pipeline velocity loss estimate from tab invisibility2x1020Quantified cost of hidden features
tech-translator → writerJargon scoring framework (1-5 scale) for all 15 tabs2x510Signals=5, Swarm=5, Projects=4, Artifacts=4 identified as worst offenders
draper → all"Ferrari with a filing cabinet bolted to the dashboard" diagnostic2x510Emotional framing that captures the core problem
debrief → conductor4 deal-room URL patterns need consolidation (/portal, /client, /room, /deals)2x510Bus-test failure: executive can't reconstruct pipeline from nav
audit-quality → conductor/caller is NOT a parallel dialer — corrected characterization3x515Prevented implementation spec from being based on incorrect architecture description

Cross-system gaps

FlaggerAffectedGapRecommended change
architectlayout.tsx/caller, /deals, /review, /atlas have no nav entry despite complete frontendsAdd nav links in layout.tsx lines 23-39
hunter/caller75 call-cards at public/projects/hrcom/call-cards/ not linked from caller dashboardWire call-cards into caller or deal detail pages
quarterback/dealsNo /deals index page exists — only [slug] sub-pagesBuild index page repurposing Admin/Dashboard's v_deal_pipeline view
listenerlayout.tsxHome page is manifesto, not dashboard — first impression is "README"Transform Home to live KPI dashboard
debriefnotebooks15 notebook MDs invisible to web UI — richest knowledge assets have no nav presenceSurface notebook entries in Swarm/Notebooks tab
architectapps/deals/171-file dead Vite/React app superseded by Next.js /deals/[slug]Delete apps/deals/
architectlaunchagentscontacts-sync plist references ~/Downloads/run_all.sh (not version-controlled)Move script to engine/ or repo root

Signal Deployment Status

SignalSupabase statusCode statusSkill/doc statusVerdict
Promote /caller to navN/AUNDONE (layout.tsx edit needed)DONE (documented in report)OPEN
Promote /deals to navN/AUNDONE (layout.tsx edit needed)DONE (documented in report)OPEN
Promote /review to navN/AUNDONE (layout.tsx edit needed)DONE (documented in report)OPEN
Build /deals index pagev_deal_pipeline view existsUNDONE (new page.tsx needed)DONE (documented in report)OPEN
Delete apps/deals/N/AUNDONE (rm -rf needed)DONE (documented in report)OPEN
Fix contacts-sync plistN/AUNDONE (move script to repo)DONE (documented in report)OPEN
Transform Home to DashboardPipeline data exists in SupabaseUNDONE (new feature build)DONE (documented in report)OPEN