Chrome DevTools Runtime Debugger

Purpose

Give agents a repeatable path for observing real browser behavior through Chrome DevTools MCP without bundling global MCP config into this plugin.

Inspect First

• local runtime URL, dev scripts, healthcheck, logs, browser test failures, console/network symptoms, and existing MCP configuration notes
• whether the user has Chrome DevTools MCP installed and available in the current session

Procedure

1. Choose the right browser tool.

   • Use Playwright tests for repeatable product workflows.
   • Use $click-path-and-browser-qa to turn repeated browser workflows into route/control inventories and Playwright coverage.
   • Use Chrome DevTools MCP for console errors, network failures, DOM/CSS inspection, Lighthouse, performance traces, responsive emulation, and live debugging.

2. Guide setup when missing.

   • Use scripts/check_chrome_devtools_mcp.py when available to check whether chrome-devtools is configured.
   • For Codex, document the official command: codex mcp add chrome-devtools -- npx chrome-devtools-mcp@latest.
   • Do not edit global Codex config unless the user explicitly asks.
   • Restart Codex or use a new session when MCP config changes are needed.

3. Run diagnosis.

   • Start local runtime and confirm healthcheck.
   • Open the target URL.
   • Inspect console and page errors.
   • Inspect failed network requests and CORS/auth/env issues.
   • Capture screenshots for layout or visual regressions.
   • Run Lighthouse or performance traces only when relevant.

4. Document findings.

   • Create or update docs/observability/browser-debugging.md from assets/templates/chrome-devtools-debugging.md.tmpl.
   • Link findings to observability-harness, testing-strategy-builder, or tdd-playwright-workflow follow-up checks.
   • Route process evidence, evaluator criteria, and task traces to sprint-contract-observability.

Validation

• Confirm the local app is reachable before browser debugging.
• Confirm Chrome DevTools MCP availability with the helper or codex mcp list; if unavailable, provide setup guidance instead of pretending the tool is callable.
• Capture concrete evidence: console message, network request, screenshot, DOM state, Lighthouse item, or trace summary.
• Confirm any fix is rechecked in the browser or covered by Playwright.

Completion Criteria

• Browser failure has concrete runtime evidence.
• Setup guidance is clear when Chrome DevTools MCP is unavailable.
• Debugging outputs are linked to tests or docs so the issue is not rediscovered from scratch.