New: 98,000+ skills indexed from GitHub repos with 100+ stars · every skill safety-checkedBrowse top skillsTop skills
Esc to closeBrowse all skills
Submit a skillBrowse skills
Submit a skillBrowse skills

Agent skill

Agent Cdp

by gronxb in gronxb/codex-relay

Chrome DevTools Protocol CLI workflow for runtime, console, network, trace, memory, and JavaScript CPU profiling analysis.

Apache-2.0Auto-check passedMobile

Install Agent Cdp

skills CLI
$ npx skills add gronxb/codex-relay --skill agent-cdp -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install gronxb/codex-relay agent-cdp --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/gronxb/codex-relay.git skills-src && mkdir -p .claude/skills && cp -r skills-src/.agents/skills/agent-cdp .claude/skills/agent-cdp && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
agent-cdp
GitHub stars
679
Token cost
~956 tokens
SKILL.md length
352 words
Files
3 (incl. references)
Skills in repo
5
Repo updated
First seen
Licence
Apache-2.0

At a glance

Chrome DevTools Protocol CLI workflow for runtime, console, network, trace, memory, and JavaScript CPU profiling analysis.

  • Works in 2 steps: Start/check the daemon → Find and select a target
  • Codex needs to inspect debuggable Chrome/Chromium tabs
  • SKILL.md covers First Step, Core Workflow, Common Tasks and Safety
  • Calls npm

What it does

Agent Cdp is an agent skill from gronxb/codex-relay. Chrome DevTools Protocol CLI workflow for runtime, console, network, trace, memory, and JavaScript CPU profiling analysis. Use when Codex needs to inspect debuggable Chrome/Chromium tabs, Node processes started with --inspect, React Native/Metro/Hermes targets, or to combine agent-device UI driving with CDP evidence for performance, heat, re-render, network, memory, or runtime-state diagnosis.

Its SKILL.md is about 960 tokens, which your agent loads only when the skill is triggered. The skill folder holds 4 other files, including reference files (for example `agents/openai.yaml` and `references/commands.md`).

It sits in Mobile, covering Cross-platform mobile apps and Browser testing. It works with React Native, Chrome DevTools and JavaScript. The repository describes itself as: Use Codex from your phone while the real work stays on your computer. The licence is Apache-2.0.

When your agent uses it

  • Codex needs to inspect debuggable Chrome/Chromium tabs
  • Node processes started with --inspect
  • React Native/Metro/Hermes targets
  • Combine agent-device UI driving with CDP evidence for performance

Example prompts

  • “/agent-cdp”

Workflow steps

2 steps, taken from the first numbered list in SKILL.md.

  1. Start/check the daemon
  2. Find and select a target

What it can do on your machine

Read from SKILL.md and the folder at commit 2a0cca2. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • npm

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md. Its commands use npm, which can reach the network depending on how they are called.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Agent Cdp loads about 956 tokens when it runs, and up to ~1.2k if it reads all its reference files. Until then it costs about 102 tokens; SKILL.md has 352 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~102
When it runs · the whole SKILL.md, loaded when a task matches
~956
With references · SKILL.md plus every file in references/, read only if the agent opens them
~1.2k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from gronxb/codex-relay at commit 2a0cca2, republished under its Apache-2.0 licence (© gronxb). 352 words, ~956 tokens.

Download SKILL.mdSave it as .claude/skills/agent-cdp/SKILL.md (or your agent's skills folder). This skill also uses 2 other files; get the full folder from GitHub.
name
agent-cdp
description
Chrome DevTools Protocol CLI workflow for runtime, console, network, trace, memory, and JavaScript CPU profiling analysis. Use when Codex needs to inspect debuggable Chrome/Chromium tabs, Node processes started with --inspect, React Native/Metro/Hermes targets, or to combine agent-device UI driving with CDP evidence for performance, heat, re-render, network, memory, or runtime-state diagnosis.

agent-cdp

Use agent-cdp for evidence-backed runtime diagnostics through Chrome DevTools Protocol. Pair it with agent-device when a mobile UI interaction must be driven on a simulator/device while CDP records what the JS runtime, network, or profiler did.

First Step

Load the bundled upstream guide before non-trivial use:

bash
agent-cdp skills get core

If that command cannot read bundled skills, fall back to the installed package files:

bash
AGENT_CDP_ROOT="$(npm root -g)/agent-cdp"
sed -n '1,220p' "$AGENT_CDP_ROOT/skills/core.md"

Core Workflow

  1. Start/check the daemon:
bash
agent-cdp start
agent-cdp status
  1. Find and select a target:
bash
agent-cdp target list
agent-cdp target list --url http://127.0.0.1:8081
agent-cdp target list --url http://127.0.0.1:9229
agent-cdp target select <target-id>

Default target scanning checks local Chrome, Node inspect, and Metro-style endpoints. If no React Native target appears, first verify Metro/dev-server reachability and whether the app is actually running in a debuggable dev mode.

React Native/Metro inspector endpoints may reject WebSocket connections with HTTP 401 unless the client sends an Origin such as http://localhost:8081. If target list sees a React Native target but target select fails with 401, verify the installed agent-cdp transport sends that Origin for targets whose kind is react-native or whose websocket URL contains /inspector/debug.

  1. Capture only the evidence needed:
bash
agent-cdp console list --limit 50
agent-cdp runtime eval --expr "globalThis.location?.href ?? process.version" --json
agent-cdp network start --name repro
agent-cdp trace start
agent-cdp profile cpu start --name hot-path
  1. Reproduce the interaction. For mobile, drive the UI with agent-device:
bash
agent-device --session <name> snapshot
agent-device --session <name> press <target>
agent-device --session <name> wait 1000
  1. Stop and inspect:
bash
agent-cdp network stop
agent-cdp network summary
agent-cdp network list --limit 20
agent-cdp trace stop
agent-cdp trace summary
agent-cdp profile cpu stop
agent-cdp profile cpu hotspots --limit 20
Show full SKILL.md (171 more words)Show less

Common Tasks

For heat or high CPU, prefer a bounded CPU profile around the suspect interaction, then inspect hotspots and stacks. Also capture a process snapshot with ps so CDP findings can be correlated with OS-level CPU.

For slow or noisy network behavior, use network start, reproduce once, then inspect network summary, filtered network list, and only specific request/response headers or bodies. Avoid dumping auth headers or secrets.

For runtime state, use runtime eval with read-only expressions. Do not mutate app state from CDP unless the user explicitly asks for that side effect.

For trace-style UI performance, use trace start/stop around a short interaction and inspect summary/tracks/entries. Export the raw trace only when compact summaries are insufficient.

For memory leaks, take baseline/action/cleanup samples or snapshots, then use the built-in memory diff/leak commands. Keep heap artifacts out of commits unless explicitly requested.

Safety

Prefer short captures over long recordings. Stop active network, trace, CPU, or allocation sessions before finishing. Report when no CDP target is available instead of pretending profile evidence exists.

© gronxb, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

SKILL.md and 2 other files (references) in .agents/skills/agent-cdp of gronxb/codex-relay.

  • SKILL.md
  • agents/openai.yaml
  • references/commands.md

Open the folder on GitHubat commit 2a0cca2

Compare with similar skills

Agent Cdp next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Agent Cdp compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Agent Cdp this skillgronxb/codex-relay679—~956Automated safety check: PassApache-2.0
Release Sample SweepAtmosphere/atmosphere3.8k—~4.2kAutomated safety check: PassApache-2.0
Crossbindcrossbind/crossbind148—~1.1kAutomated safety check: PassMIT
Argent Metro Debuggerbbplayer-app/BBPlayer1.1k—~3.4kAutomated safety check: PassMIT
Multithreadingkingstinct/react-native-healthkit7161 repos~1.1kAutomated safety check: PassMIT
Diagnosing Stacktrace SymbolicationPostHog/posthog40k—~2.2kAutomated safety check: PassCustom licence

Similar skills

  • Release Sample Sweep

    Atmosphere/atmosphere

    Run the pre-release end-to-end sweep of every user-facing surface — the 33 samples under samples/ (booted from their packaged artifacts and driven in a real browser via chrome-devtools MCP), the…

    3.8k GitHub stars~4.2k tokensUpdated 2 days ago
    MobileAuto-check passed

  • Crossbind

    crossbind/crossbind

    A skill your agent uses when a user wants to call C++ or Rust from JavaScript or TypeScript; add a native library such as GDAL, SQLite, OpenSSL, GEOS or PROJ to a browser, Node.js, Cloudflare Worker…

    148 GitHub stars~1.1k tokensUpdated today
    MobileAuto-check passed

  • Argent Metro Debugger

    bbplayer-app/BBPlayer

    Debug a JS runtime via CDP using argent debugger tools. An agent skill from bbplayer-app/BBPlayer.

    1.1k GitHub stars~3.4k tokensUpdated today
    MobileAuto-check passed

  • Multithreading

    kingstinct/react-native-healthkit

    Software Mansion's best practices for multithreading in React Native apps using react-native-worklets.

    716 GitHub starsUsed in 1 repo~1.1k tokens
    MobileAuto-check passed

  • Diagnosing Stacktrace Symbolication

    PostHog/posthog

    Official

    Help users debug PostHog Error Tracking stack-trace symbolication for any supported platform — JavaScript/TypeScript web, React Native (Hermes), Android (Proguard / R8), or iOS / macOS (dSYM).

    40k GitHub stars~2.2k tokensUpdated today
    MobileAuto-check passed

  • API Design

    margelo/react-native-skills

    Design and review predictable public APIs for TypeScript, JavaScript, React, and React Native libraries.

    175 GitHub stars~7.2k tokensUpdated 1 mo ago
    MobileAuto-check passed

More from gronxb/codex-relay

  • Hot Updater

    gronxb/codex-relay

    Set up, upgrade, operate, and diagnose Hot Updater projects.

    679 GitHub stars~2.7k tokensUpdated yesterday
    Auto-check passed

  • Hot Updater QA Ship

    gronxb/codex-relay

    Deploy the current Modex iOS app version with Hot Updater for QA-only delivery.

    679 GitHub stars~818 tokensUpdated yesterday
    Auto-check passed

  • Agent Device

    gronxb/codex-relay

    Automates interactions for Apple-platform apps (iOS, tvOS, macOS) and Android devices.

    679 GitHub stars~1.8k tokensUpdated yesterday
    Auto-check passed

  • Deploy iOS

    gronxb/codex-relay

    Deploy the Modex/Codex Relay mobile app iOS OTA update with Hot Updater.

    679 GitHub stars~382 tokensUpdated yesterday
    Auto-check passed

Works with

Categories

Questions about Agent Cdp

What does Agent Cdp do?

Chrome DevTools Protocol CLI workflow for runtime, console, network, trace, memory, and JavaScript CPU profiling analysis. Agent Cdp is an agent skill from gronxb/codex-relay. Chrome DevTools Protocol CLI workflow for runtime, console, network, trace, memory, and JavaScript CPU profiling analysis.

When should I use Agent Cdp?

Agent Cdp fits situations like: Codex needs to inspect debuggable Chrome/Chromium tabs; Node processes started with --inspect; React Native/Metro/Hermes targets; combine agent-device UI driving with CDP evidence for performance.

How do I install Agent Cdp in Claude Code?

Run `npx skills add gronxb/codex-relay --skill agent-cdp -a claude-code`. Or copy the skill folder (.agents/skills/agent-cdp in gronxb/codex-relay) into .claude/skills/agent-cdp in your project. Claude Code loads it when a task matches its description.

How do I install Agent Cdp in Codex?

Run `npx skills add gronxb/codex-relay --skill agent-cdp -a codex`. Or copy the skill folder (.agents/skills/agent-cdp in gronxb/codex-relay) into .agents/skills/agent-cdp in your project. Codex loads it when a task matches its description.

Can I use Agent Cdp in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add gronxb/codex-relay --skill agent-cdp -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/agent-cdp, .gemini/skills/agent-cdp, .github/skills/agent-cdp and .opencode/skills/agent-cdp in your project.

What does Agent Cdp need to run?

Going by SKILL.md and its folder, Agent Cdp needs the command-line tools its instructions call (npm).

Does Agent Cdp access the network?

SKILL.md contains no URLs. Its commands use npm, which can reach the network depending on how they are called. This is read from the text; nothing was executed.

Is Agent Cdp safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Agent Cdp use?

Agent Cdp is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Agent Cdp use?

About 956 tokens (SKILL.md is roughly 3.8k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 238 tokens, read only when the agent opens those files.

What are the alternatives to Agent Cdp?

Skills that share tags, products or a category with Agent Cdp: Release Sample Sweep (Atmosphere/atmosphere, 3.8k stars), Crossbind (crossbind/crossbind, 148 stars), Argent Metro Debugger (bbplayer-app/BBPlayer, 1.1k stars) and Multithreading (kingstinct/react-native-healthkit, 716 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Agent Cdp?

gronxb (a GitHub user) maintains it in gronxb/codex-relay, which has 679 GitHub stars. The repository holds 5 skills in this directory. The repository was last updated on October 7, 2026.

Source: gronxb/codex-relay on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.