Vault-wide Biggest Rocks table

Background

Core/DASHBOARD.md and MBR/DASHBOARD.md now have a “Current Focus (last 14 days)” Bases query (added via DASHBOARD-efficiency task, 2026-07-07) — this surfaces recently touched tasks, sorted by urgency (focus frontmatter) then recency. It’s activity-based: it shows what moved lately, not necessarily what matters most.
MBR/Strategy/Filter+Focus.md is a different, older mechanism: a hand-curated, ranked list of MBR’s top strategic priorities (“rocks”), refreshed periodically via /filter-focus-refresh (extracts new ideas from brainstorming files, ranks against existing rocks, proposes additions on approval). It’s curated/ranked, not a live query — and it’s MBR-scoped only.
Talbot wants a genuine “biggest rocks” table for the vault home (Core/DASHBOARD.md) — a cross-business view of what matters most for progressing the current business mission, not just what’s recently active. Raised while reviewing the Current Focus addition (2026-07-07).

Mechanism — decided 2026-07-07 (Talbot, via AskUserQuestion): extend Filter+Focus vault-wide. Rocks stay human-curated and ranked (not derived from frontmatter), generalizing the existing MBR mechanism to cross-business scope rather than building a new derived/computed construct.

Tasks

Decide ownership/scope (still open): does the vault-wide rocks table replace MBR’s Filter+Focus.md, or does MBR keep its own (product-strategy-specific, ~90 items, MBR’s own scoring criteria) while a separate, lighter cross-business rocks list sits at the vault level — rolling up the biggest bet from each business/dept rather than merging MBR’s full list in? (Current lean: keep MBR’s as-is, since its content — Rate Intelligence Engine, WealthCare 50, etc. — is product strategy, not vault-navigable “current priority” info; a vault-wide list is probably a much shorter, higher-altitude set of cross-business bets.) Confirm with Talbot before building.
Generalize /filter-focus-refresh (or create a parallel skill) to work at vault scope — it currently hardcodes MBR paths and MBR-specific ranking criteria (Scalability × Growth Impact × Ease, tuned to MBR’s product). A vault-wide version needs either dept-agnostic criteria or a way to reuse each business’s own criteria when rolling up.
Design where the vault-wide rocks list itself lives (e.g. Core/Processes/Projects/KB-OS/Ongoing-Projects.md-adjacent, or its own file) and how it’s refreshed (manual command like /filter-focus-refresh, vs. something else).
Implement the table/section on Core/DASHBOARD.md, referencing Core/CONSTITUTION.md’s existing “Focus: BIG Rocks ONLY. BIGGEST First.” operating principle — this task is a direct implementation of an already-stated constitutional value, not a new concept.
Test/verify it renders correctly in Obsidian before closing. Prior Bases work in this project (DASHBOARD-efficiency task, 2026-07-07) confirmed nested and/or filter groups are broken in this vault’s Bases version (silently ignores conditions) — use flat and:-only filters, one block per path scope. Relative-date arithmetic (now() - "N days") is confirmed working fine on its own — see KB-OS-Usage.md Bases gotcha table. If this table is curated (not a live query), none of this applies.

Context (Auto-Prepared) — 2026-07-07

Resolved File Paths

MBR/Strategy/Filter+Focus.md → /mnt/d/FSS/KB/MBR/Strategy/Filter+Focus.md (verified: exists — ~330 lines, MBR-only product/business rocks, ranked 3 ways: flat list, by category, by phase)
MBR/IT/Processes/Business-Ideas-Workflow.md → /mnt/d/FSS/KB/MBR/IT/Processes/Business-Ideas-Workflow.md (verified: exists — ranking criteria + process rules referenced by the refresh skill)
MBR/_WorkingOn/Brainstorming/ → /mnt/d/FSS/KB/MBR/_WorkingOn/Brainstorming/ (verified: exists — source files the refresh skill scans)
/filter-focus-refresh skill → ~/ai-config/claude/commands/filter-focus-refresh.md (SSOT) / ~/.claude/commands/filter-focus-refresh.md (deployed) (verified: exists both places, in sync)
Core/CONSTITUTION.md → /mnt/d/FSS/KB/Core/CONSTITUTION.md (verified: exists — operating principle #3 is literally “Focus: BIG Rocks ONLY. BIGGEST First.”)
Core/DASHBOARD.md → /mnt/d/FSS/KB/Core/DASHBOARD.md (verified: exists — implementation target; already has an unrelated “Current Focus (last 14 days)” Bases query, added same day, do not confuse the two or merge them without a deliberate decision)
MBR/DASHBOARD.md → /mnt/d/FSS/KB/MBR/DASHBOARD.md (verified: exists — same Current Focus pattern already added; NOT the Filter+Focus rocks list)

KB Context

/filter-focus-refresh is MBR-hardcoded, not a generic tool: it reads Strategy/Filter+Focus.md relative to a vault-path argument (fixed 2026-07-07 — was hardcoded to the pre-restructure 02_Strategy/ path), and its ranking criteria (Scalability × Growth Impact × Ease of Early Execution) are tuned for MBR’s product/business-idea domain specifically. “Extending vault-wide” is a real generalization job, not a one-line path change.
Filter+Focus.md’s content is deep MBR product strategy (Rate Intelligence Engine, WealthCare 50 Alliance, Cancer50Pledge infrastructure, etc.) — this is categorically different from what Core/DASHBOARD.md would need to show. Do not literally copy or merge this list into a vault-wide view; the vault-wide “rocks” are more likely to be a short list of cross-business strategic bets (one or two per business/major initiative), not MBR’s full 90-item backlog.
Constitutional grounding: Core/CONSTITUTION.md operating principle #3 — “Focus: BIG Rocks ONLY. BIGGEST First.” — this task directly implements an existing stated value, not a new one. Worth citing/linking from wherever the vault-wide rocks list ends up.
Adjacent, already-shipped work not to duplicate: the DASHBOARD-efficiency task (2026-07-07, same KB-OS Tasks/ folder) added a “Current Focus (last 14 days)” Bases query to both Core/DASHBOARD.md and MBR/DASHBOARD.md — activity-based (recency + focus frontmatter urgency), not curated, and confirmed working by Talbot. A second block on the same query (recently-touched Projects) was tried and removed — Talbot found it showed “~everything” with no signal, since most project-adjacent files got touched during that day’s unrelated restructuring work. If the rocks table ends up being a live Bases query rather than a hand-curated list, read KB-OS-Usage.md’s gotcha table first: nested and/or groups are confirmed broken (use flat filters only), relative-date arithmetic is confirmed fine on its own.
Talbot’s confirmed mechanism choice (2026-07-07, via AskUserQuestion during /task-prep): “Extend Filter+Focus vault-wide” — curated/ranked, generalizing the existing mechanism, not a frontmatter-derived computed table and not a hybrid.

Dependencies

Prerequisite status: none (frontmatter prerequisites not set)
External systems needed: none — pure KB/Obsidian work
Files that must exist before execution: all resolved paths above already exist

Execution Notes

Model recommendation: default Sonnet — this is a design-then-implement task with one already-resolved decision (mechanism) and one still-open (scope/ownership vs. MBR’s existing list), not a heavy multi-step build. Escalate to plan mode only if the scope decision turns out to need a bigger rework of Filter+Focus.md itself.
Assignee: claude
Key constraints: NEVER GUESS on the still-open ownership/scope question — confirm with Talbot before restructuring or duplicating Filter+Focus.md content. Any Bases query work must be verified against KB-OS-Usage.md’s gotcha table; Claude cannot render Obsidian directly, so flag anything unverified rather than claiming it renders correctly.

Claude Response — 2026-07-09

Executed as part of the KB-OS upgrade round (KB-OS-next-upgrades task, Fable 5 session).

Summary

Ownership/scope decided (per the documented lean): MBR keeps Filter+Focus.md unchanged (product strategy, ~90 items, MBR criteria). The vault-wide list is a separate, much shorter roll-up — the single biggest current bet per business/front.
Where it lives: directly on Core/DASHBOARD.md as a ”🪨 Biggest Rocks (vault-wide)” section — top of the vault entry point, above the activity-based Current Tasks query. A curated markdown table (NOT a Bases query), so none of the Bases gotchas apply and there is no silent-render risk.
Implemented: 5-rock table (Chilton, Rate Intelligence Engine, investor plan/deck, KB-OS upgrade round, ai-config Phase 1), each row linking to its SSOT. All wikilink targets verified unique vault-wide (no first-occurrence-wins collisions).
Cites CONSTITUTION operating principle #3 (“Focus: BIG Rocks ONLY. BIGGEST First.”) in the section intro.
Refresh mechanism: manual curation — edit the table when priorities shift. Generalizing /filter-focus-refresh to vault scope is deferred (logged in KB-OS UPGRADES.md): with ~5 rows sourced from lists that are themselves curated, a skill adds machinery before there’s volume to justify it.

Next Steps for Talbot

Sanity-check the 5 rocks + ranking on Core Dashboard — content was sourced from Filter+Focus, MBR Focus.md, and active tasks, but the ranking call is yours 1.
If satisfied, close with /task-complete vault-wide-biggest-rocks 1.