← Back to Skills Marketplace
ivangdavila

Brave Browser

by Iván · GitHub ↗ · v1.0.0 · MIT-0
linuxdarwinwin32 ✓ Security Clean
493
Downloads
1
Stars
3
Active Installs
1
Versions
Install in OpenClaw
/install brave
Description
Operate, automate, and troubleshoot Brave Browser with profiles, Shields, extensions, and Chromium debugging workflows.
README (SKILL.md)

When to Use

User needs help with Brave Browser itself, not generic browsing advice. Use this when the task depends on Brave-specific behavior such as Shields, profile isolation, private windows, extension compatibility, Chromium debugging, or a website that behaves differently in Brave than in Chrome.

Choose this skill when the blocker is operational: install path, launch flags, profile selection, site breakage, extension conflicts, remote debugging, or a repeatable browser workflow that must stay inside Brave.

Architecture

Memory lives in ~/brave/. If ~/brave/ does not exist, run setup.md. See memory-template.md for structure.

~/brave/
|-- memory.md          # Durable activation rules, OS facts, and safety boundaries
|-- profiles.md        # Known profiles, purpose, and launch notes
|-- sites.md           # Site-specific Shields overrides and known-good fixes
|-- automation.md      # Remote-debugging ports, test flows, and tool preferences
`-- incidents.md       # Startup failures, extension conflicts, and recovery notes

Quick Reference

Load only the smallest file needed for the current blocker.

Topic File
Setup guide setup.md
Memory template memory-template.md
Launch commands and profile strategy launch-and-profiles.md
Shields and site-breakage recovery shields-and-compatibility.md
Brave automation and DevTools workflows automation-and-debugging.md
Extensions, sync, wallet, and private windows extensions-and-wallet.md
Failure recovery and cleanup order troubleshooting.md

Requirements

  • Brave Browser should already be installed, or there must be permission to guide installation.
  • Verify the operating system, install path, and target profile before changing anything.
  • Ask for explicit approval before enabling remote debugging, clearing profile data, importing wallet or sync state, changing global Shields defaults, or opening multiple sensitive sites.
  • Treat login sessions, cookies, private windows, sync, and wallet data as user-sensitive state.

Operating Coverage

This skill is for operating Brave as a browser platform, not for generic web search or Brave Search API setup. It covers:

  • app launch and path detection across macOS, Windows, and Linux
  • normal profiles, disposable profiles, and private-window workflows
  • site failures caused by Shields, aggressive blocking, or extension conflicts
  • Chromium-compatible automation using Brave with Playwright, Puppeteer, or remote debugging
  • extension compatibility, sync boundaries, and wallet-adjacent caution
  • browser recovery when startup, updates, or profile state go wrong

Data Storage

Keep only durable Brave operating context in ~/brave/:

  • approved profiles and what each one is for
  • known site fixes and whether they are per-site or global
  • allowed automation posture, remote-debugging defaults, and no-go actions
  • repeated incidents such as crash-on-launch, bad flags, corrupted profile state, or extension collisions

Core Rules

1. Identify the Exact Brave Surface Before Acting

  • Lock four facts first: OS, install path, target profile, and whether the issue is launch, browsing, extension, or automation related.
  • Brave problems often look like "the browser is broken" when the real cause is the wrong profile or a stale flag.
  • Do not change settings until the target surface is explicit.

2. Treat Shields as the First Compatibility Check

  • When a site loops on login, breaks scripts, hides media, or blocks checkout, inspect Brave Shields before assuming the site itself is bad.
  • Prefer per-site fixes before global relaxations so privacy defaults stay intact elsewhere.
  • Use shields-and-compatibility.md to move from symptom to the smallest reversible change.

3. Keep Profiles Separate by Risk and Purpose

  • Use one stable daily profile and separate profiles for testing, automation, or risky extensions.
  • Do not debug a broken production login inside the same profile used for experiments.
  • If a fresh profile fixes the issue, record that before touching the main profile.

4. Reuse Chromium Tooling, But Verify Brave-Specific Behavior

  • Brave supports Chromium-style automation, but Shields, built-in blockers, and profile choices can change outcomes.
  • Use automation-and-debugging.md when attaching Playwright, Puppeteer, or a DevTools client.
  • A workflow that passes in Chrome is not proof that it will pass in Brave without adjustment.

5. Treat Extensions, Sync, Wallet, and Private Windows as Trust Boundaries

  • Extension installs, sync enablement, wallet access, and private-window workflows affect sensitive user state.
  • Ask before changing permissions, importing state, or enabling anything that broadens data exposure.
  • Never present private windows or Tor-based browsing as a way to bypass site restrictions or anti-fraud controls.

6. Change One Variable at a Time

  • When diagnosing, isolate profile, Shields, extension set, launch flag, or remote-debugging state instead of changing several together.
  • One controlled change plus one verification step beats "try everything" browser debugging.
  • Record the exact change that fixed the issue so the user does not repeat the incident.

7. Verify in the Browser, Not Only in the Command Output

  • A launch command succeeding does not prove the right window, profile, or site state is live.
  • Confirm the actual browser state: expected profile opened, extension loaded, site behavior changed, or DevTools endpoint became reachable.
  • If the expected state is not visible, stop and switch to a safer fallback.

Brave Traps

  • Treating Brave as "just Chrome with a different icon" -> Shields and privacy defaults keep surprising the workflow.
  • Disabling blockers globally for one broken site -> privacy posture degrades everywhere.
  • Testing automation in the user's daily profile -> cookies, extensions, and session state skew the result.
  • Clearing profile data too early -> the hardest-to-recover state gets destroyed first.
  • Mixing sync, wallet, and private-window issues in one troubleshooting pass -> the blast radius becomes unclear.
  • Assuming a site failure is anti-bot hostility -> many breaks are local compatibility problems and can be fixed reversibly.

Security & Privacy

Data that may leave your machine:

  • normal website traffic to the domains the user opens in Brave
  • optional sync, extension-store, or wallet-related traffic only if the user already uses those Brave features
  • optional automation traffic to a local DevTools endpoint when the user enables remote debugging

Data that stays local:

  • browser state already stored by Brave in its own profile directories
  • durable operating notes under ~/brave/ if the user approves persistence

This skill does NOT:

  • use undeclared remote APIs by default
  • recommend bypassing bot checks, paywalls, or fraud controls
  • assume remote debugging is safe to leave on permanently
  • clear profiles or security-sensitive browser state without explicit approval

Scope

This skill ONLY:

  • helps operate Brave Browser safely and predictably
  • structures profile, Shields, extension, and automation work into reversible steps
  • keeps durable notes for approved profiles, site fixes, and recurring incidents

This skill NEVER:

  • act as a generic search-engine skill
  • require the user to adopt sync, wallet, or private-window workflows they did not ask for
  • store secrets, passwords, or full browsing history in its own memory files
  • modify its own skill files

Related Skills

Install with clawhub install \x3Cslug> if user confirms:

  • chrome - Reuse Chromium debugging patterns when the issue is not Brave-specific.
  • playwright - Automate and verify web flows after the Brave launch surface is stable.
  • puppeteer - Drive DevTools and Chromium automation with lower-level script control.
  • macos - Handle macOS app paths, permissions, and automation details around Brave on Apple systems.
  • web - Fetch or inspect web content after the browser environment is behaving correctly.

Feedback

  • If useful: clawhub star brave
  • Stay updated: clawhub sync
Usage Guidance
This skill appears to be a careful, instruction-only operator for Brave Browser and is coherent with its described purpose. Before using it, confirm you are comfortable with the skill creating and updating files under ~/brave/ (it states it will ask for consent). Also approve explicitly before enabling remote debugging, making persistent automation changes, or allowing any wallet/sync modifications. If you see requests for unrelated credentials, network endpoints, or instructions to copy secrets into notes, decline and stop the operation.
Capability Analysis
Type: OpenClaw Skill Name: brave Version: 1.0.0 The 'brave' skill bundle is a well-structured set of instructions for managing and automating the Brave Browser. It emphasizes user privacy, explicit consent for sensitive actions (such as remote debugging or clearing profile data), and safe troubleshooting practices. The skill includes specific safeguards against mishandling sensitive data like crypto wallets or seed phrases and focuses entirely on legitimate browser operations without any indicators of malicious intent, data exfiltration, or unauthorized execution.
Capability Assessment
Purpose & Capability
Name/description (Brave Browser operator) matches the content: launch paths, profiles, Shields, automation/DevTools, extensions, and troubleshooting. The requested config path (~/brave/) is proportional for storing operating context and playbooks.
Instruction Scope
SKILL.md and supporting docs stay within Brave-specific operational scope. Instructions explicitly require user approval before enabling remote debugging, clearing profile data, touching sync/wallet, or making global Shields changes. They direct storage to a single declared local path and do not instruct reading unrelated system files or exfiltrating data.
Install Mechanism
No install spec and no code files — instruction-only — so nothing is downloaded or written automatically during install. This minimizes disk-write and remote-code risks.
Credentials
No environment variables or credentials are requested. The single declared config path (~/brave/) is appropriate for durable skill memory; the docs require user consent before writing persistent state and explicitly avoid requesting secrets (e.g., wallet seed phrases).
Persistence & Privilege
Skill is not always-enabled and is user-invocable. It asks to store durable context under ~/brave/ only with consent. There is no indication it modifies other skills or system-wide agent settings.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install brave
  3. After installation, invoke the skill by name or use /brave
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
Initial release with profile workflows, site-compatibility diagnostics, extension handling, and browser recovery playbooks.
Metadata
Slug brave
Version 1.0.0
License MIT-0
All-time Installs 3
Active Installs 3
Total Versions 1
Frequently Asked Questions

What is Brave Browser?

Operate, automate, and troubleshoot Brave Browser with profiles, Shields, extensions, and Chromium debugging workflows. It is an AI Agent Skill for Claude Code / OpenClaw, with 493 downloads so far.

How do I install Brave Browser?

Run "/install brave" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Brave Browser free?

Yes, Brave Browser is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Brave Browser support?

Brave Browser is cross-platform and runs anywhere OpenClaw / Claude Code is available (linux, darwin, win32).

Who created Brave Browser?

It is built and maintained by Iván (@ivangdavila); the current version is v1.0.0.

💬 Comments