Build executable safety net tests

Generate executable characterization tests from prototype behavior

Locking down current prototype behavior with executable tests before enterprise replatforming to prevent regression.

• · Running prototype (localhost URL or live server)
• · DISCOVERY_REPORT.md from vibe-browser-audit
• · Browser flow recordings and network traces (from runtime-observer)

• · Characterization test files (Jest, Python pytest, or language-appropriate)
• · Static JSON fixtures under tests/characterization/fixtures/
• · Fixture portability report (confirms no secrets/paths leak)
• · Runtime telemetry report with latency baselines

• · runtime-observer agent (captures browser/network telemetry)
• · Playwright or Puppeteer for CDP browser automation
• · JSON fixture scrubber for portability validation

• · Prototype running and accessible at known URL
• · DISCOVERY_REPORT.md completed by vibe-browser-audit
• · runtime-observer configured for payload/timing capture

• · Fixture contains real API key: portability check fails, scrub and retry
• · Test uses network endpoints instead of mocks: refactor to load fixtures locally
• · Timing baseline too loose: test becomes fragile, tighten based on observed latency

• · Tests load static JSON fixtures, not live endpoints
• · Fixture portability report confirms zero secrets exposure
• · Each test preserves exact legacy behavior (even quirks) as safety net
• · Timing baselines documented for performance regression detection