feat(scaling): back ceilings and progress with runs

[0x00101010]

Summary

Why

The first screen's execution ceiling, qualified rung and progress entries are typed in by hand, with no link to the runs that would back them. This wires them to bench runs and load tests so every number names its run, and an estimate always says why it is one.

What

• Execution ceiling card: fetches bench metadata in the browser, takes the newest passing benchmark run (the fastest passing configuration when it tested several), shows its sequencer Mgas/s tagged Estimated with "state size not recorded" and a link to the run. The snapshot value stays when the API is unavailable or when the snapshot already has a measured or modeled value. The estimate never becomes the binding ceiling.
• Qualified rung: a value now requires ladder.qualified.evidence links to /scaling/runs/... pages (validator), and the ladder lists them.
• Area pages: "Related runs" shows the newest five bench runs and five load tests covering the area, linking to the full index.
• Progress: an evidence.url that points at a run page shows that run's Mgas/s and status inline.
• The run table and index loader move out of app/scaling/runs/page.tsx unchanged so area pages reuse them.

Notes

• Stack: feat(scaling): snapshot schema v2 with ladder, ceilings and guardrails #194 → feat(scaling): first screen with ladder, ceilings, guardrails and progress #196 → feat(scaling): detail pages and unified run index #199 → feat(scaling): bench run and load test detail pages #203 → this PR.
• No producer records state size yet, so the card cannot show Measured; that needs base-bench to record it.

Screenshots

Internal dev build with a synthetic snapshot (the committed fixture plus a run-backed qualified rung and work evidence) and report-api responses replaced by the committed synthetic run fixture.

/scaling at 1280 px and 360 px (execution card from the run, qualified rung links, run status in Progress):

/scaling/execution at 1280 px with related runs, and /scaling/state (no run covers state):

Report-api unreachable: the cards fall back to the snapshot and related runs shows the error.

Test plan

• npm test, npm run typecheck, npm run lint, npm run build, npm run build:internal, npm run llms -- --check, npm run agents -- --check
• runBacked.test.tsx: estimate with reason and link, never binding below a modeled DA ceiling; snapshot fallback when the API fails; measured snapshot kept; qualified links; related runs per area and kind; run status in Progress.
• snapshot.test.ts: a qualified value without run-page evidence fails validation.
• Rendered /scaling and the area pages in dev:internal against live report-api (execution card shows the newest live bench run, Estimated), the synthetic fixtures, and with the API unreachable, at 360 px and 1280 px.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Actions	Updated
omni-ui	Ready	Preview	Oct 16, 2026 2:04pm UTC

[cb-heimdall]

✅ Heimdall Review Status

Requirement	Status	More Info
Reviews	✅ 1/1	Denominator calculation Show calculation 1 if user is bot 0 1 if user is external 0 2 if repo is sensitive 0 From .codeflow.yml 1 Additional review requirements Show calculation Max 0 0 From CODEOWNERS 0 Global minimum 0 Max 1 1 1 if commit is unverified 0 Sum 1