thinkbrowse-mcp

A browser-control skill for ThinkBrowse MCP, a tool server that lets an agent operate a web browser. It can navigate pages, inspect their structure, interact with controls, extract content, and capture screenshots.

In plain words
What is it for?
Use it to open websites, inspect page content, click buttons, enter text, take screenshots, and record browser sessions through ThinkBrowse MCP tools.
Why use it?
It provides a defined way to work with browser pages when the ThinkBrowse or ThinkRun MCP server is explicitly available. Reading the page structure before acting helps the agent choose the right elements.

Skill for Claude CodeCodex

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/dundas/thinkrun/thinkbrowse-mcp
Any agent
npx skills add dundas/thinkrun --skill thinkbrowse-mcp
Clone the repo
git clone --depth 1 https://github.com/dundas/thinkrun

Made for: Claude Code, Codex.

Per session 81 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,600 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
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 $0.00081 $0.01600
Opus 5 $0.00041 $0.00800
Sonnet 5 $0.00016 $0.00320
Haiku 4.5 $0.00008 $0.00160

Measured yesterday against content hash 556045fbab65, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

thinkbrowse-mcp 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 yesterday.

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.

.claude/skills/thinkbrowse-mcp/SKILL.md · 183 lines

How it starts

The opening of the file, as written. The whole thing — 183 lines — stays where its author put it; the contents beside it link to each section on GitHub.

ThinkBrowse MCP

Control browsers using ThinkBrowse MCP tools (mcp__thinkbrowse__*). 35 tools. Three modes: auto (default), local, cloud.

Every tool call MUST include a thought parameter explaining your reasoning.

Quick Start

1. session_create          → get a session
2. session_wait_ready      → wait for provisioning (cloud)
3. navigate                → go to URL
4. snapshot                → read page structure (best for AI)
5. screenshot              → visual capture
6. click / type_text       → interact
7. session_delete          → clean up

Top 5 Patterns

1. Navigate → Snapshot → Act

snapshot  {thought: "Read page structure before interacting"}
click  {selector: "button[type=submit]", thought: "Click submit — confirmed via snapshot"}

Always snapshot before clicking to find the right selector.

2. Click by Text (When Selectors Are Ambiguous)

evaluate  {
  script: "[...document.querySelectorAll('button,a,[role=button]')].find(el=>el.textContent.trim()==='Sign In')?.click()",
  thought: "Multiple buttons — clicking by visible text"
}

3. React Forms (type_text, NOT fill)

click  {selector: "input[name=email]", thought: "Focus email field"}
type_text  {selector: "input[name=email]", text: "[email protected]", thought: "React input — using type_text"}
click  {selector: "button[type=submit]", thought: "Submit form"}

fill uses native DOM value setting — React won't detect the change. type_text uses keyboard events.

4. List Interactive Elements

evaluate  {
  script: "[...document.querySelectorAll('button,[role=tab],a')].map(b=>({label:b.getAttribute('aria-label'),text:b.textContent.trim().slice(0,50),tag:b.tagName})).filter(b=>b.text||b.label)",
  thought: "Find all clickable elements"
}

5. Cloud Session Lifecycle

session_create  {thought: "Create cloud session"}
session_wait_ready  {sessionId: "...", thought: "Wait for provisioning (60-90s)"}
navigate  {url: "https://example.com", thought: "Go to target"}
...
session_delete  {sessionId: "...", thought: "Clean up"}

Read the full file on GitHub · 183 lines

Files

What ships with it

4 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

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. yesterday First seen · 183 lines · 81 tokens per session scan A 556045fbab65

Subscribe to this mod's changes

thinkbrowse-mcp is a skill published in the GitHub repository dundas/thinkrun (1 stars, last pushed 1mo ago), licensed MIT. It adds 81 tokens to every session and 1,600 once invoked, about $0.0004 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 2026-08-31.

Related

Other skills, from other repositories

Agent Browser Automation

Fast Rust-based headless browser automation CLI with Node.js fallback for AI agents, featuring navigation, clicking, typing, snapshots, and structured commands optimized for agent workflows.

PramodDutta/qaskills · 37 tokens

browseweave

Use when installing, repairing, verifying, diagnosing, or safely operating the BrowseWeave systemd-based Linux developer preview with Chrome, Zen, or Firefox and a local MCP client. Do not use for store publication or bypassing site protections.

xenitV1/browseweave · 52 tokens

wcag22-a11y-audit

WCAG 2.2 Accessibility Audit skill that systematically evaluates web pages against 8 core Success Criteria (1.1.1, 1.4.3, 1.4.11, 2.1.1, 2.1.2, 2.4.3, 2.4.7, 4.1.2) using accessibility tree inspection and visual analysis. Use this skill when you need to perform accessibility testing/auditing on a live webpage.

AIPexStudio/AIPex · 111 tokens

skill-creator

Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends AIPex's capabilities with specialized knowledge, workflows, or tool integrations.

AIPexStudio/AIPex · 47 tokens

ux-audit-walkthrough

Minimalist UX/Interaction Audit Expert that deconstructs complex interactions through cognitive load and operational efficiency lenses. Use this skill when you need to perform a UX walkthrough audit on a Figma prototype or web interface, evaluating usability based on principles like fewer clicks, less UI elements, no…

AIPexStudio/AIPex · 73 tokens

Angry User Simulator

Simulate aggressive user behavior patterns including rapid clicking, random navigation, form abuse, tab spamming, and unexpected interaction sequences to find UI resilience issues.

PramodDutta/qaskills · 34 tokens