browser-companion

browser-companion is a skill for Claude Code from joaquimscosta/arkhe-claude-plugins. It costs 138 tokens per session (1,252 once invoked), scanned A, original, MIT.

A local preview server that displays HTML files in a browser and reloads the page when they change. It also records browser clicks so the agent can read what the user selected.

In plain words
What is it for?
It helps preview HTML fragments, compare design variants, run selection flows, capture click events, and stop the preview server when finished.
Why use it?
It removes the need to repeatedly refresh a browser or describe clicks manually during design and mockup work.

Skill for Claude Code

Written for Claude Code: disable-model-invocation in frontmatter.

Runs only inside its plugin — its command needs a path that Claude Code sets for a plugin’s own hooks and for nothing else. Install the plugin, not this.

Part of the devtools plugin — 9 skills shipped together

Good fit It helps preview HTML fragments, compare design variants, run selection flows, capture click events, and stop the preview server when finished.

Compare 6 skills from other repositories ↓
Install

Getting it into your agent

This one installs as part of its plugin. Adding the marketplace and installing the plugin brings it with everything else the plugin ships.

Claude Code
/plugin marketplace add joaquimscosta/arkhe-claude-plugins
Claude Code
/plugin install devtools

Made for: Claude Code.

Or install devtools, the plugin that ships this one along with the rest of its 9 skills.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for browser-companion

README.md
[![agentmods](https://agentmods.dev/badge/skills/joaquimscosta/arkhe-claude-plugins/browser-companion/github.svg)](https://agentmods.dev/skills/joaquimscosta/arkhe-claude-plugins/browser-companion)
Your own site
<a href="https://agentmods.dev/skills/joaquimscosta/arkhe-claude-plugins/browser-companion"><img src="https://agentmods.dev/badge/skills/joaquimscosta/arkhe-claude-plugins/browser-companion/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for browser-companion

Your own site · 80×15
<a href="https://agentmods.dev/skills/joaquimscosta/arkhe-claude-plugins/browser-companion"><img src="https://agentmods.dev/badge/skills/joaquimscosta/arkhe-claude-plugins/browser-companion.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 138 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,252 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00138 $0.01252
Opus 5 $0.00069 $0.00626
Sonnet 5 $0.00028 $0.00250
Haiku 4.5 $0.00014 $0.00125

Measured 11d ago against content hash cb963059e156, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

browser-companion scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 11d ago.

The scan reads SKILL.md. This mod also ships 7 executable files (examples/brainstorm/helper.js, examples/gallery/helper.js, scripts/helper.js, …), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/devtools/skills/browser-companion/SKILL.md · 113 lines

How it starts

The opening of the file, as written. The whole thing — 113 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Browser Companion

A zero-dependency Node.js HTTP + WebSocket server that turns a directory of HTML fragments into a live-reloading preview, with browser→agent click event capture.

The skill ships a public CLI arkhe-preview placed in plugins/devtools/bin/. Any plugin's skill can invoke it as a bare bash command — no need to know the devtools install path.

Quick Start

# Start a server scoped to the current project
arkhe-preview start --project-dir "$(pwd)"

# Output (one JSON line on stdout):
#   {"type":"server-started","port":54123,"url":"http://localhost:54123",
#    "screen_dir":".../<id>/content","state_dir":".../<id>/state",
#    "session_dir":".../<id>","pid":12345}

# Tell the user to open the URL. Then write HTML to screen_dir:
echo '<h2>Hello</h2><button data-event="ok">OK</button>' > "$SCREEN/content.html"

# Browser auto-reloads. User clicks. Read events:
tail -f "$STATE/events.jsonl"

# When done:
arkhe-preview stop "$SESSION_DIR"

When to Use This

  • Design / mockup loops — write HTML to screen_dir, browser shows it
  • Multi-select / variant selection — give the user clickable options, read events.jsonl to learn what they picked
  • Iteration on generated UI — overwrite fragment → instant reload
  • Cross-plugin — any plugin's skill can call arkhe-preview directly; it lives in devtools but its CLI is on PATH whenever devtools is enabled

Core Commands

arkhe-preview start [options]      # Print {url, screen_dir, state_dir, session_dir, pid} JSON
arkhe-preview stop <session-dir>   # Graceful SIGTERM, falls back to SIGKILL
arkhe-preview status <session-dir> # {running, pid, url} JSON (always exits 0)
arkhe-preview --help               # Full option reference (USAGE.txt)

Common start options:

  • --project-dir <path> — store under <path>/.claude/preview/<id>/ (default: /tmp/arkhe-preview-<id>/)
  • --frame-template <path> — override default HTML frame
  • --helper <path> — override default browser-side helper JS
  • --port <N> — pin a specific port
  • --owner-pid 0 — disable the watchdog (use idle timeout only)

Read the full file on GitHub · 113 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 11d ago First seen · 113 lines · 138 tokens per session scan A cb963059e156

Subscribe to this mod's changes

browser-companion is a skill published in the GitHub repository joaquimscosta/arkhe-claude-plugins (21 stars, last pushed 28d ago), licensed MIT. It adds 138 tokens to every session and 1,252 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

demo-video

Orchestrates building a narrated demo video of a project — reads the codebase, writes a storyboard, prepares deterministic app state, drives the UI with Playwright to record clips (web and Electron), generates ElevenLabs voiceover, reconciles measured durations into a timeline, and renders the final cut with Remotion.…

lukaskellerstein/claude-my-marketplace · 116 tokens

demo-capture

Records demo video clips by driving a running app — one clip per storyboard section, with human-feeling pointer motion, dwell timing, and per-character typing. Covers the Playwright MCP video tools for web apps and a generated Playwright Electron script (plus a macOS screen-capture fallback) for desktop apps. Use when…

lukaskellerstein/claude-my-marketplace · 91 tokens

browser-edge-cases

SOP for debugging browser automation failures on complex websites. Use when browser tools fail on specific sites like LinkedIn, Twitter/X, SPAs, or sites with Shadow DOM.

aden-hive/hive · 40 tokens

memstack-development-webapp-testing

Use when the user says 'write browser tests', 'test this page', 'playwright test', 'e2e test', 'end to end test', 'browser test', 'test the UI', or needs Playwright-based browser testing for a web application. Do NOT use for unit tests, API tests, or non-browser testing.

cwinvestments/memstack · 75 tokens

web-access

A method for handling difficult web tasks across search, page fetching, and a real browser. It covers sites that need login sessions, dynamic pages, anti-bot handling, or exploration before their structure is understood.

open-octo/octo-agent · 139 tokens

responsiveness-check

Test website responsiveness across viewport widths using browser automation. Resizes a single session through breakpoints, screenshots each width, and detects layout transitions (column changes, nav switches, overflow). Produces comparison reports showing exactly where layouts break. Trigger with 'responsiveness…

jezweb/claude-skills · 85 tokens