Chrome Devtools MCP AI Context Pack for Claude Code, Codex, Cursor, and Aider

Pack version: v1.0.4 · Last updated: 2026-05-14

Languages: English | 中文

Stop guessing browser bugs. Give your AI coding agent a DevTools evidence loop: inspect console/network signals, name the failure mode, and recover from common MCP setup failures without pretending the tool is already installed.

This is an independent AI context resource pack for ChromeDevTools/chrome-devtools-mcp: host instructions, prompt preview, evals, pitfalls, and recovery rules you can load into Claude Code, Codex, Cursor, Aider, and other AI coding agents.

This is an independent capability pack. It is not affiliated with or endorsed by ChromeDevTools/chrome-devtools-mcp unless explicitly stated.

Copy / Run / Verify

1. Copy AGENTS.md or CLAUDE.md into your AI coding host.
2. Run the first prompt in 01_PROMPT_PREVIEW.md.
3. Verify behavior with 06_EVALS/smoke_check.md, then recover with 03_PITFALL_LOG.md if it fails.

Quick links: Start · Prompt · Evals · Pitfalls · Manual

When This Helps

Use this pack when you want an AI coding agent to understand the upstream project context, setup boundaries, common pitfalls, and verification steps before it edits files, installs dependencies, or claims the tool is ready.

What You Get

• Host instructions for AI coding agents.
• A copyable prompt preview.
• Smoke, boundary, and failure checks.
• Pitfall log and recovery steps.
• Boundary and risk card.
• Human reference manual (05_HUMAN_MANUAL.md) — architecture, components, and failure taxonomy.
• Source attribution and upstream links.

If this pack helps your agent work from evidence instead of guesses, star the repo so future updates are easier to find. Open an issue for bugs, usage questions, or new pitfall reports.

AGENTS.md for Claude Code and AI Coding Agents

Use AGENTS.md when your coding host supports repository instructions. Use CLAUDE.md when Claude Code is the target host. Both files tell the agent to inspect DevTools evidence before claiming browser behavior is fixed.

Chrome DevTools MCP Pitfalls and Recovery

Start with 03_PITFALL_LOG.md when setup, permissions, runtime behavior, or verification fails. The pack separates host-level eval success from real Chrome DevTools MCP runtime install success.

Source Attribution

This project pack was assembled by Doramagic to make ChromeDevTools/chrome-devtools-mcp usable as a portable AI capability asset.

• Upstream/source: https://github.com/ChromeDevTools/chrome-devtools-mcp
• License: Apache-2.0
• Pack contents: prompts, host instructions, checks, guardrails, and validation notes
• Relationship: independent pack; not affiliated with or endorsed by ChromeDevTools/chrome-devtools-mcp unless explicitly stated

If you maintain the upstream project and want attribution changed or removed, open an issue in this repository.