Add mobench 0.1.28 BrowserStack benchmark pipeline

[dcbuild3r]

Summary

• add a bench-mobile workspace crate for the monolithic passport complete_age_check flow with prepare, prove, verify, and e2e mobench benchmarks
• embed deterministic Noir fixtures for noir-examples/noir-passport-monolithic/complete_age_check and add local smoke tests for prove/verify
• add BrowserStack mobile benchmark workflows, a sticky PR comment flow using <!-- mobench-summary -->, and a local mobench patch for device discovery / markdown rendering
• benchmark across three Android and three iOS BrowserStack devices using low, mid, and flagship tiers

Device matrix

• Android low: Google Pixel 6-12.0
• Android mid: Google Pixel 7-13.0
• Android flagship: Samsung Galaxy S24-14.0
• iOS low: iPhone 13-15
• iOS mid: iPhone 14-16
• iOS flagship: iPhone 16 Pro-18

Validation

• cargo fmt --all --check
• cargo test -p bench-mobile --test passport_smoke
• cargo-mobench list --crate-path ./bench-mobile
• actionlint .github/workflows/mobile-bench.yml .github/workflows/mobile-bench-pr-auto.yml .github/workflows/mobile-bench-pr-command.yml .github/workflows/mobile-bench-reusable.yml

Not measured here

• live BrowserStack runs were not executed locally from this environment
• sticky comment rendering and PR-triggered mobile runs depend on repo Actions + BrowserStack secrets in GitHub

[dcbuild3r]

Mobench Benchmark Results

Status: waiting for the first BrowserStack benchmark run.

Configured device matrix:

• Android: Google Pixel 6-12.0, Google Pixel 7-13.0, Samsung Galaxy S24-14.0
• iOS: iPhone 13-15, iPhone 14-16, iPhone 16 Pro-18

This comment is intentionally seeded with the mobench marker so the workflow can update it in place once the benchmark job runs.

---

Posted by mobench seed setup at $(date -u '+%Y-%m-%d %H:%M UTC')

[dcbuild3r]

Superseded by #398, which uses the same branch from worldfnd/provekit so BrowserStack benchmarks and sticky comments run in the correct repo context.