feat(stats): add accuracy and reuse-rate sparklines to the homepage hero (#4447, #4448)

[JSONbored]

Summary

• Adds an 8-week sparkline to the existing Decision accuracy tile and a new fifth AI work reused tile (also sparklined) in the homepage's ProofOfPowerStats hero band, riding the accuracyTrend / reuseRateTrend series GET /v1/public/stats already returns (feat(review): add a public weekly review-accuracy trend (#4447) #4655, feat(review): add cache hit/miss telemetry to grounding and review-memory (#4448) #4658, feat(review): instrument impact-map + add a public AI-reuse-rate trend (#4448) #4671).
• This was originally scoped as a standalone /stats page with full trend charts, then explicitly redirected to inline hero sparklines instead — smaller surface, no new route, no navigation change.
• A week below its own minimum-sample floor renders as accuracyPct: null / reuseRatePct: null; the sparkline draws a genuine gap there (connectNulls={false}) instead of a fabricated flat line or a zero.
• No backend/OpenAPI changes — this is a pure consumer of fields already shipped and already covered by codecov/patch in their own PRs, so this diff has no src/** lines and no coverage obligation.

Scope

• The PR title follows type(scope): short summary Conventional Commit format, for example fix(api): restore profile access checks.
• This PR is focused and does not mix unrelated backend, UI, MCP, docs, dependency, and deploy changes.
• This follows CONTRIBUTING.md and does not reintroduce GitHub Pages, VitePress, site/, or CNAME.
• I linked a currently open issue this PR resolves (e.g. Closes #123) — a linked open issue is required for every contributor PR.

Closes #4448 (last unchecked deliverable — "Homepage//stats page shows the reuse-rate trend" — the other #4448 deliverables shipped in #4658/#4671; the "daily rollup, cron-refreshed" deliverable was deliberately built as a live-computed query over durable tables instead, mirroring the same live-vs-rollup call made for #4447, to avoid a cron-miss gap and a second copy of the number). Advances #4447 (already closed; this adds the "accuracy-over-time chart alongside the existing live number" piece its own checklist left undone at close time).

Validation

• git diff --check
• npm run actionlint — skipped, no .github/workflows/** changes in this diff.
• npm run typecheck
• npm run test:coverage locally; codecov/patch requires ≥99% coverage of the lines AND branches you changed (aim for 100% on your diff so CI variance does not fail near the threshold). Global coverage is a non-blocking trend with a loose 90% backstop, not the gate. — skipped: no src/** lines in this diff (UI-only), so codecov/patch doesn't apply. UI-side behavior is covered by ui:test instead (see below).
• npm run test:workers — skipped, no Worker/backend code touched.
• npm run build:mcp — skipped, no MCP package changes.
• npm run test:mcp-pack — skipped, no MCP package changes.
• npm run ui:openapi:check — clean; this diff doesn't touch any route or schema.
• npm run ui:lint
• npm run ui:typecheck
• npm run ui:build
• npm audit --audit-level=moderate
• New or changed behavior has unit/integration tests for new branches, fallback paths, and sanitizer boundaries — proof-of-power-stats.test.tsx: five-tile assertion (was four), a dedicated sparkline-count/a11y-name test, and the fixture now carries a deliberate null-accuracy week to exercise the gap path.

Also ran, not in the template's list: npm --workspace @jsonbored/gittensory-ui run test (15 files / 113 tests, full workspace) both before and after rebasing onto latest main — no drift.

Safety

• No secrets, wallet details, hotkeys, coldkeys, user PATs, private keys, raw trust scores, private rankings, or private maintainer evidence are exposed.
• Public GitHub text stays sanitized, low-noise, and does not imply compensation guarantees or optimization tactics.
• Auth, cookie, CORS, GitHub App, Cloudflare, or session changes include negative-path tests. — N/A, no auth/session/CORS code touched.
• API/OpenAPI/MCP behavior is updated and tested where needed. — N/A, no API/OpenAPI/MCP changes; pure frontend consumer of already-shipped fields.
• UI changes use live API data or real empty/error/loading states, not production mock/demo fallbacks — both tiles read straight off the live /v1/public/stats payload; the whole section still renders nothing until totals.handled > 0, same as before.
• Visible UI changes include a UI Evidence section below with JPG/JPEG or PNG screenshots arranged as organized, captioned, clickable thumbnails. SVG screenshots are not used as review evidence. Review-only screenshots or recordings are not committed to the repository. — left unchecked honestly: see UI Evidence below for what was actually verified and why no GitHub-hosted image URLs are attached.
• Public docs/changelogs are updated where needed; changelogs are only edited for release-prep PRs — no changelog edit (not a release-prep PR).

UI Evidence

No GitHub-hosted screenshot URLs are attached — this session has no browser access to github.com to produce user-attachments/user-images URLs, and the template asks not to commit review-evidence images to the repo. In lieu of that, here's exactly what was verified against the local dev server (apps/gittensory-ui, via Claude Code's Preview tooling, real recharts render in a real DOM, not a snapshot test):

State	What was checked
Loaded, dark (default)	All 5 tiles render; both sparklines draw with stroke: oklch(0.88 0.18 115) (--chart-1, accuracy) and oklch(0.75 0.16 140) (--chart-2, reuse rate) — the app's own existing, already-validated chart ramp, not a new palette.
Loaded, light	Toggled the .dark class off directly (this app keys theming off a class, not prefers-color-scheme) — both sparklines correctly repaint to their light-mode OKLCH values (0.78 0.18 115 / light chart-2), no hardcoded dark-only color.
Null-week gap	Inspected the accuracy sparkline's rendered SVG <path> directly: two M (moveto) subpaths, breaking exactly at the fixture's null-accuracy week — confirms connectNulls={false} produces a real visual gap for "below minimum sample," not an interpolated line through it.
Mobile (375×812)	grid-cols-2 layout: the 5th tile ("AI work reused") wraps to its own row at half width as expected, sparkline stays proportioned inside the tile, no overflow/clipping.
Mark spec	Both <Line>s: strokeWidth: 2px, strokeLinecap/strokeLinejoin: round, no axes/gridlines/tooltip (sparklines are the documented exception to the hover-layer default) — matches the dataviz skill's stat-tile figure contract.

Notes

• .claude/launch.json (a local Preview-tooling dev-server config) is intentionally not included in this PR — it's outside wantedPaths and is personal/session tooling, not a repo need.

[cloudflare-workers-and-pages]

Deploying with    Cloudflare Workers

The latest updates on your project. Learn more about integrating Git with Workers.

Status	Name	Latest Commit	Preview URL	Updated (UTC)
✅ Deployment successful! View logs	gittensory-ui	10cf1f0	Commit Preview URL Branch Preview URL	Jul 10 2026, 11:56 AM

[superagent-security]

Superagent didn't find any vulnerabilities or security issues in this PR.