Build in public logTracker line

feat/product-drift-sample-demo (advisor fix merged in): the recorder's DATA_DIR data/demo is a demo-named dir, so HIGHFI_DRIFT_SAMPLE=demo still serves the rows unmarked there; the same server on a dir not named demo serves them marked with the tag (lane proof both ways, snapshot into a scratch folder opened as a demo build: no tag, no console error); no row, line, state or count changed; no recorder run, nothing under site/demo.

Time
2026-10-11 09:09
Where it was logged
context-log/living/05-TRACKER-LOG.md, line 557
Session: highfi-product-fidelity
Brought to main by
d9e6e4465f Merge pull request #220 from lenilsonjr/develop (2026-10-13 11:36)The rows below that say "with it" describe this commit, which can carry other work from the same pull request.
Impact on decisions

Names no decision record.

Records created or changed with it:

  • created: D-100 With GitHub Actions back, the queue ships into develop on the real CI checks; local mode only when CI is unavailable
  • created: D-101 Superseded branches are deleted on origin so they cannot be queued by mistake
  • created: D-102 The Otto M4 port checkpoint branch is retired; M6.2 and later Otto work build on develop's otto.pen
  • created: D-103 The records branch's rulebook delta was accepted on the core lead's behalf while the core lead was paused
  • created: D-104 Path B (Otto's real code compared) moves out of demo-projects-complete into its own roadmap entry
  • created: D-105 The deep amethyst logo mark is dropped
  • created: D-106 The facilitator accepted the dashboard Fable review branch while the services lead was paused
  • created: D-107 The unit coverage report becomes a gate with floors of 65 lines, 72 branches and 65 functions
  • created: D-108 server/ gets a coverage target of 70 percent of lines, non-blocking
  • created: D-109 The need about the core lead's pending prompt in pane w5:pE is dropped as stale
  • created: D-110 The published dashboard republishes on every ship
  • created: D-111 Fable tokens are spent on the dashboard review
  • created: D-112 Proposed: the Sandbox/HighFi/HighFi Context shell is archived, not deleted
  • created: D-113 Copying otto-core/.highfi into otto-frontend-v2 waits
  • created: D-114 The Otto and GoBookIt seed declarations stay in HighFi
  • created: D-115 D-79's metadata and component promotion apply to design/highfi.pen and design/otto.pen only
  • created: D-116 The Otto React components live in design/components/otto/react/, beside the HTML replica
  • created: D-117 The P-03 Amended section is accepted as written
  • created: D-118 One writer per pen file is made checkable with a claim file
  • created: D-119 The responsive percentage counts destinations with a drawn phone or wide state
  • created: D-120 The codebase review umbrella closes on S1 to S3
  • created: D-121 Merged highfi/* branches auto-delete after 30 days; unmerged ones need a person
  • created: D-122 Agent runs keep the shared node_modules, with a per-run lockfile check
  • created: D-123 Faint text passes the contrast check only where it carries information
  • created: D-124 Drift at 4,155 entries on a phone: first paint under 1.5 seconds and 50 fps scroll
  • created: D-125 The company is formed through Stripe Atlas this week, and the two founders hold the signing key
  • created: D-126 Nic is the first outside user, as a first customer, and gets HighFi for free
  • created: D-127 The unit coverage floor of 65 lines, 72 branches and 65 functions, confirmed by Devin
  • created: D-128 One real driver run is made end to end, and only Devin sets the flags in .env.highfi.local
  • created: D-129 Otto Marketplace's inputs do not run on GitHub runners; the Otto demo is not recorded in CI
  • created: D-130 Five questions Devin deferred on the help desk stay open: s17, s20, h1, h6 and p4
  • created: D-131 Switch stage 5 is the core lead's top priority, because fidelity on design/highfi.pen cannot be measured before it lands
  • created: D-132 The same-size compare (product branch 0) was started on the facilitator's recommendation, ahead of Devin's Q3 answer
  • created: D-133 Devin's fidelity answers: both halves at 100 or a written reason, Otto on placeholder code, the Pen frame's box, the visitor path first, a walkthrough page
  • created: D-134 Three stage 5 decisions are accepted for the core lead: the maps flip with the file, the Library reference is the frame NrgL1 for now, the dogfood specs move with stage 5
  • created: D-135 Fidelity work is batched two to three families per branch, and a family that is not ready leaves the batch
  • created: D-136 Fidelity rules A to E: tokens before families, a pilot batch of atoms, visitor-path frames frozen for the day, one records pass per hour, noise gets a written reason
  • created: D-137 The design session's two asks are granted in their later form: the catalog leaves product.pen only in the stage 6 archive, and the otto-core re-exports wait for stage 5
  • created: D-138 The compare shows 'not comparable: the Pen export is a board card' before stage 5's baselines, merged into stage 5 ahead of track D
  • created: D-139 The order after stage 5: the fonts fix, then map-visitor re-pointed to plain names, then design's master re-exports
  • created: D-140 Otto design work pauses in HighFi; the Otto atomic system moves to Otto Group
  • created: D-142 HighFi answers on this computer only unless the operator turns the network on (HIGHFI_LAN=1); amends D-04 and D-78
  • created: D-143 A five-day framework library spike runs before the component library decision is sealed; no record changes until its findings
  • changed: D-87 One hashing authority: the Rust engine writes and checks the locks, the app reads them, and hifi check gates CI
  • changed: D-95 Two atomic .pen files in the HighFi repo, HighFi and Otto, and their exact replicas
  • created: D-96 develop as the integration branch while CI is unavailable
  • created: D-97 Pen design work runs headless through the pen CLI, pinned at 0.3.10; one renderer per measurement
  • created: D-98 The develop into main pull request (#220) opened under the u1 default and holds for Lenilson's review
  • created: D-99 Models are set by role; no lane runs on Fable by default
  • changed: P-02 Pricing and licensing (non-commercial 49, commercial 499 per seat, team of three 1,499; first 500 commercial at 199, lifetime)
  • created: P-11 The beta starter is a ShadCN code library with its HighFi file; no design file in the starter for the MVP
  • created: P-12 A records pass per merge
Scope expanded or reduced

Expanded: roadmap entries added

  • Dialogs, canvas and touch targets (100 percent)
  • Accessibility and visual regression checks (96 percent)
  • Alignment framework and dashboard (100 percent)
  • Atomic design system (HighFi) and its exact replica (67 percent)
  • Brand identity and logo (96 percent)
  • Code health and tests of the prototype (48 percent)
  • Delivery engine (land, clean, ship without stalls) (88 percent)
  • Demo project complete (HighFi) (100 percent)
  • Design coverage: frames the product still lacks (100 percent)
  • Event loop and the drift job (60 percent)
  • Fable review with the bounded prompts (100 percent)
  • Fixtures: actors, entities, packs, scenes (84 percent)
  • Git safety and driver stub coverage (92 percent)
  • JavaScript CI and test isolation (100 percent)
  • Records and the conductor (README, context log, roadmap, plans, decisions) (89 percent)
  • Server core unit tests (85 percent)
  • Shell and form factor (100 percent)
  • Site polish and copy after the critique round (100 percent)

Reduced: roadmap entries removed

  • Dialogs, canvas and touch targets
  • Accessibility and visual regression checks
  • Alignment framework and dashboard
  • Atomic design systems (HighFi and Otto) and their exact replicas
  • Brand identity and logo
  • Code health and tests of the prototype
  • Delivery engine (land, clean, ship without stalls)
  • Demo projects complete (HighFi and Otto Marketplace)
  • Design coverage: frames the product still lacks
  • Event loop and the drift job
  • Fixtures: actors, entities, packs, scenes
  • Git safety and driver stub coverage
  • JavaScript CI and test isolation
  • Records and the conductor (README, context log, roadmap, plans, decisions)
  • Server core unit tests
  • Shell and form factor
  • Site polish and copy after the critique round
Plan clarity
  • MVP: a team onboards its own codebase in seven days: output clarity 70 to 72 (raised)
Milestones and metrics: roadmap entries moved

No roadmap entry's percentage moved with it.

Superseded
not recorded (tracker lines carry no supersedes field)