Skip to content
FSS Knowledge Base

Focus — Upgrades

Section titled “Focus — Upgrades”

Ideas discovered but not committed to a phase. Promote into ROADMAP.md when concrete.

  • Comet pinned-tab switch (superseded, see ROADMAP): Talbot found no “Arguments” field in the shortcut’s Properties dialog (Windows .lnk only exposes a single Target string — args must be typed inline after the exe path, e.g. "comet.exe" --app=<url>, which launches a second instance/window rather than acting on the running one). “On startup → Open a specific page” was tried next and rejected — it discards all previously-open tabs. Current fix: “Continue where you left off” + a pinned Focus tab, tracked in ROADMAP.md.
  • filter-focus-refresh cadence: no defined trigger yet for when to run the skill — candidate for a scheduled/cron nudge once Focus project has more than one active thread.
  • Focus-page loader duplication grew, not shrunk. A 2026-06-25 monorepo architecture review already flagged sites/mbr/src/utils/load-focus-content.ts and sites/sdc/src/utils/load-focus-content.ts as copy-pasted (severity: high, ~2h to extract to a shared package). The Mission-SSOT embed work (2026-07-30) added ~65 more near-identical lines to both files (resolveEmbeds, convertWikilinks, extractSection, escapeRegex). Consolidating into a shared @monorepo/focus-content (or similar) package is now a bigger win than it was — same 2h-ish estimate, larger payoff. Not done as part of this task (scope was the SSOT fix, not a refactor) — candidate for ROADMAP.md when picked up.
  • Ecco Pro replacement — two named gaps, not a general search (2026-08-20). Talbot’s stated “ideal focus/task app” (one app · local + private · no vendor lock-in · agent context per business/dept · task context and logs · model-agnostic) already describes KB-OS + SMTM. Only two things keep Ecco alive: (1) hierarchical collapsible outlining, (2) any-device access. Framing the search as those two gaps is far cheaper than “find the one app”, and either could be closed without leaving the vault (Obsidian outlining plugins; Obsidian Sync or a deployed read-only view for mobile). Alternatives Talbot has been tracking live inside Notes.eco. Do not start a build here on enthusiasm alone — that is the “build custom where an existing tool would do” pattern named in SDC/Strategy/Identity/Internal View/Weaknesses.md.
  • Bridge the Focus Coach into Ecco (offered, not built, 2026-08-19). The SMTM gates only fire inside Claude Code sessions; Talbot’s day starts on the handwritten My Focus page and Ecco’s Today view. Cheapest possible bridge: reuse Ecco’s existing Focus column and set its values to the current Biggest Rock names, so both systems speak the same vocabulary. Manual, ~5 minutes at each monthly Rocks review — no integration, no code.
  • Focus-ledger review has no teeth yet (2026-08-19). /task-complete now appends a row per closed task, and the monthly Routines.md row says to read it — but nothing acts on what it shows. Revisit after ~10 rows: if the off-rock share is high, the answer is probably re-ranking the Rocks (they are wrong), not a harder gate.
  • MBR/Strategy/Filter+Focus.md is stale — date-refreshed: 2026-04-08 (2026-08-20). Core/DASHBOARD.md’s Rocks header names it as the business’s full ranked list that the vault-wide roll-up summarises, so the roll-up currently points at a list nobody has touched in 4.5 months. Either refresh it (/filter-focus-refresh) at a Friday planning block, or demote it in the header to a historical artifact — but it should not keep the title of live SSOT while stale.
  • Dataview Serializer for ecco-view.js output — approved, pending install (2026-08-25). ecco-view.js’s tree renders only exist at Obsidian render time — invisible to grep/AI reading the raw file. Dataview Serializer materializes query results as real Markdown between comment markers, kept in sync on save/command — a plausible fix. Talbot approved the install 2026-08-25 (revertible if it doesn’t pan out). Tracked for install + fit test in Core/Processes/Projects/KB-OS/Tasks/Obsidian-Upgrades-Phase2-Verification.md.
  • Kanban Action Planner — free, WBS hierarchy view — approved, pending install (2026-08-25). github.com/dsebastien/obsidian-kanban-action-planner, MIT license, works standalone (no Starter Kit dependency, confirmed at source). Its WBS view renders parent/child hierarchy as a collapsible tree — goal → projects → tasks — with rollup calculations for estimates/progress. This is the exact capability ecco-replacement-research (2026-08-20) said no candidate had: Bases is note-granularity (flattens hierarchy), TaskNotes shares that shape. Talbot approved the install+test 2026-08-25 — this is the load-bearing test for whether Ecco stays irreplaceable. Desktop-only (no mobile support yet, per its docs). Tracked in Core/Processes/Projects/KB-OS/Tasks/Obsidian-Upgrades-Phase2-Verification.md; report the result back here once tested.
  • Related-tasks/related-notes Bases pattern (2026-08-25, from “No Plugins, No AI” clipping). Native-Obsidian Bases queries (no plugin) that surface tasks/notes related to whatever file is open — linked, backlinked, name-match, or 2-tag overlap. Worth trying before TaskNotes: it may deliver “tasks discoverable from any context” without TaskNotes’ one-note-per-task shape (already rejected for flattening hierarchy in ecco-replacement-research). See KB-OS UPGRADES.md for the query snippets.
  • Notify desktop app overlaps the alert layer (2026-08-20). notify (desktop text + speech) and Google Calendar + BuzzKill (phone, unmissable) both answer “tell me now”. BuzzKill reaches Talbot anywhere; notify only reaches the desk. Worth checking at some point whether notify still earns its place for focus work, or is now only the Achilles-rehab tool. No action yet — usage data first.