Agent Browser CLI
vercel-labs/agent-browser
Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking…
Guidelines for driving a workspace's embedded browser with the fewest tool calls, smallest state reads and no needless screenshots or waits.
Install Efficient Embedded Browser Loop
$ npx skills add holaboss-ai/holaOS --skill browser-core-efficient -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install holaboss-ai/holaOS browser-core-efficient --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/holaboss-ai/holaOS.git skills-src && mkdir -p .claude/skills && cp -r skills-src/runtime/harnesses/src/embedded-skills/browser-core-efficient .claude/skills/browser-core-efficient && rm -rf skills-srcUse ~/.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/
Install the "browser-core-efficient" agent skill from https://github.com/holaboss-ai/holaOS/tree/main/runtime/harnesses/src/embedded-skills/browser-core-efficient into .claude/skills/browser-core-efficient/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "browser-core-efficient", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/holaboss-ai/holaOS/tree/main/runtime/harnesses/src/embedded-skills/browser-core-efficientType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add holaboss-ai/holaOS --skill browser-core-efficient -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install holaboss-ai/holaOS browser-core-efficient --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/holaboss-ai/holaOS.git skills-src && mkdir -p .agents/skills && cp -r skills-src/runtime/harnesses/src/embedded-skills/browser-core-efficient .agents/skills/browser-core-efficient && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "browser-core-efficient" agent skill from https://github.com/holaboss-ai/holaOS/tree/main/runtime/harnesses/src/embedded-skills/browser-core-efficient into .agents/skills/browser-core-efficient/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "browser-core-efficient", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add holaboss-ai/holaOS --skill browser-core-efficient -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install holaboss-ai/holaOS browser-core-efficient --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/holaboss-ai/holaOS.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/runtime/harnesses/src/embedded-skills/browser-core-efficient .cursor/skills/browser-core-efficient && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "browser-core-efficient" agent skill from https://github.com/holaboss-ai/holaOS/tree/main/runtime/harnesses/src/embedded-skills/browser-core-efficient into .cursor/skills/browser-core-efficient/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "browser-core-efficient", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/holaboss-ai/holaOS.git --path runtime/harnesses/src/embedded-skills/browser-core-efficient--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add holaboss-ai/holaOS --skill browser-core-efficient -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install holaboss-ai/holaOS browser-core-efficient --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/holaboss-ai/holaOS.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/runtime/harnesses/src/embedded-skills/browser-core-efficient .gemini/skills/browser-core-efficient && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "browser-core-efficient" agent skill from https://github.com/holaboss-ai/holaOS/tree/main/runtime/harnesses/src/embedded-skills/browser-core-efficient into .gemini/skills/browser-core-efficient/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "browser-core-efficient", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install holaboss-ai/holaOS browser-core-efficientInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add holaboss-ai/holaOS --skill browser-core-efficient -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/holaboss-ai/holaOS.git skills-src && mkdir -p .github/skills && cp -r skills-src/runtime/harnesses/src/embedded-skills/browser-core-efficient .github/skills/browser-core-efficient && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "browser-core-efficient" agent skill from https://github.com/holaboss-ai/holaOS/tree/main/runtime/harnesses/src/embedded-skills/browser-core-efficient into .github/skills/browser-core-efficient/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "browser-core-efficient", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add holaboss-ai/holaOS --skill browser-core-efficient -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install holaboss-ai/holaOS browser-core-efficient --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/holaboss-ai/holaOS.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/runtime/harnesses/src/embedded-skills/browser-core-efficient .opencode/skills/browser-core-efficient && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "browser-core-efficient" agent skill from https://github.com/holaboss-ai/holaOS/tree/main/runtime/harnesses/src/embedded-skills/browser-core-efficient into .opencode/skills/browser-core-efficient/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "browser-core-efficient", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
Facts
- Skill name
browser-core-efficient- GitHub stars
- 11k
- Token cost
- ~736 tokens
- SKILL.md length
- 377 words
- Files
- 1
- Skills in repo
- 30
- Repo updated
- First seen
- Licence
- Custom licence
At a glance
Guidelines for driving a workspace's embedded browser with the fewest tool calls, smallest state reads and no needless screenshots or waits.
- Works in 5 steps: If the target is already known, use… → If the target is semantically known but… → Use browser_get_state only for… → …
- Automating a task in the workspace's embedded browser with few tool calls
- SKILL.md covers Goals, Default Loop, Tool Selection and State Discipline, plus 2 more sections
- Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md
What it does
The default loop is to act directly when the target is known, using `browser_act` with locator signals such as text, label, role, selector or ref; to use `browser_find` when you know what you want on a large page but not where it is; and to call `browser_get_state` only for orientation. Index-based `browser_click` and `browser_type` are reserved for a stable index from a fresh state snapshot, and tab tools replace extra re-orienting calls.
State is kept small: compact detail by default, no page text unless reading content is the task, and no screenshots unless DOM signals fall short. When a fact may sit in custom-element attributes, `data-*` attributes, hrefs or hydration data, the agent inspects those before concluding it is unavailable, and treats refs as stale after a page change. For waiting, it prefers `wait_for` on the action itself, with `interactive` or `domcontentloaded` for light transitions, full load only when needed, and `download_started` and `download_completed` for downloads. Standard detail and page text are escalations, not defaults.
When your agent uses it
- Automating a task in the workspace's embedded browser with few tool calls
- Locating an element on a large page without dumping the whole state
- Waiting for navigation or a download without polling
Example prompts
- “Open the pricing page in the embedded browser and click the Enterprise plan button using as few calls as possible.”
- “Find the search box on this long page and type the query without reading the full page text.”
- “Start the report download and wait until it finishes.”
Requirements
- The workspace's embedded browser tools, such as browser_act and browser_find
Workflow steps
5 steps, taken from the first numbered list in SKILL.md.
- If the target is already known, use browser_act directly with locator signals.
- If the target is semantically known but the page is large, use browser_find.
- Use browser_get_state only for orientation or when the target is still unclear.
- After page-changing actions, use wait_for on the action itself when possible.
- Re-read state only if the action result is ambiguous.
What it can do on your machine
Read from SKILL.md and the folder at commit 4684714. 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
No scripts in the folder and no shell commands in SKILL.md.
From the folder's file list and the shell code blocks in SKILL.md.
Network
No URLs in SKILL.md.
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
Efficient Embedded Browser Loop loads about 736 tokens when it runs. Until then it costs about 29 tokens; SKILL.md has 377 words of instructions outside code blocks.
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
Its licence (Custom licence) doesn't allow us to republish the file, so here is its outline and opening line. It has 377 words (~736 tokens).
“Use this skill when operating the workspace-embedded browser and efficiency matters.”
- name
- browser-core-efficient
Files
Just SKILL.md in runtime/harnesses/src/embedded-skills/browser-core-efficient of holaboss-ai/holaOS.
Open the folder on GitHubat commit 4684714
Compare with similar skills
Efficient Embedded Browser Loop 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.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Efficient Embedded Browser Loop this skillholaboss-ai/holaOS | 11k | — | ~736 | Automated safety check: Pass | Custom licence | |
| Agent Browser CLIvercel-labs/agent-browser | 44k | 24 repos | ~864 | Automated safety check: Pass | Apache-2.0 | |
| Agent Browserquran/quran.com-frontend-next | 1.9k | 41 repos | ~3.3k | Automated safety check: Pass | None | |
| Web Access via Browser CDPeze-is/web-access | 9.1k | 4 repos | ~2.2k | Automated safety check: Pass | MIT | |
| Dev Browser AutomationMemTensor/MemOS | 12k | 3 repos | ~1.7k | Automated safety check: Pass | Apache-2.0 | |
| Electron App Automationvercel-labs/agent-browser | 44k | 5 repos | ~1.7k | Automated safety check: Pass | Apache-2.0 |
Similar skills
- Official
Agent Browser
quran/quran.com-frontend-next
Automates browser interactions for web testing, form filling, screenshots, and data extraction.
Web Access via Browser CDP
eze-is/web-access
Routes every web task, from searching to logged-in browsing, through a tiered choice of search, fetch, curl or a real Chrome or Edge session driven over CDP.
Dev Browser Automation
MemTensor/MemOS
Automates a real browser through short TypeScript scripts that keep page state between runs, for navigating, filling forms, taking screenshots and extracting data.
- Official
Electron App Automation
vercel-labs/agent-browser
Automates Electron desktop apps such as VS Code, Slack or Discord by connecting agent-browser to their Chrome DevTools Protocol port.
- Official
Slack Browser Automation
vercel-labs/agent-browser
Drives the Slack web app with the agent-browser CLI to check unread channels, search, read channel details and extract information, with screenshots as evidence.
More from holaboss-ai/holaOS
All 30 skills in this repoInterface Design for Dashboards and Apps
holaboss-ai/holaOS
Pushes an agent past generic defaults when designing dashboards, admin panels, SaaS apps and tools, with attention to structure, type, navigation and how data is shown.
Customize an Installed Capability
holaboss-ai/holaOS
Edits an already-installed workspace capability's manifest to add an MCP server, a skill, an integration or agent guidance, then reinstalls it so the change takes effect on the next run.
Dashboard Builder for holaOS
holaboss-ai/holaOS
Builds the visual layer of a holaOS dashboard app with TanStack Start and @holaboss/ui, starting from a bundled reference instead of a blank page.
Capability Creator
holaboss-ai/holaOS
Bundles one or more existing skills and the integrations they need into a single named, installable capability.
Browser QA Repro Loop
holaboss-ai/holaOS
Validates or debugs a workflow in the embedded browser with a reproducible, evidence-first loop that separates product failures from wait or locator problems.
holaOS MCP Configurator
holaboss-ai/holaOS
Adds, removes or updates MCP servers in a holaOS workspace by editing workspace.yaml in the mcp_registry format, with rules for headers, local servers and tool allowlists.
Related
Categories
Questions about Efficient Embedded Browser Loop
What does Efficient Embedded Browser Loop do?
Guidelines for driving a workspace's embedded browser with the fewest tool calls, smallest state reads and no needless screenshots or waits. The default loop is to act directly when the target is known, using `browser_act` with locator signals such as text, label, role, selector or ref; to use `browser_find` when you know what you want on a large page but not where it is; and to call `browser_get_state` only for orientation. Index-based `browser_click` and `browser_type` are reserved for a stable index from a fresh state snapshot, and tab tools replace extra re-orienting calls.
When should I use Efficient Embedded Browser Loop?
Efficient Embedded Browser Loop fits situations like: automating a task in the workspace's embedded browser with few tool calls; locating an element on a large page without dumping the whole state; waiting for navigation or a download without polling.
How do I install Efficient Embedded Browser Loop in Claude Code?
Run `npx skills add holaboss-ai/holaOS --skill browser-core-efficient -a claude-code`. Or copy the skill folder (runtime/harnesses/src/embedded-skills/browser-core-efficient in holaboss-ai/holaOS) into .claude/skills/browser-core-efficient in your project. Claude Code loads it when a task matches its description.
How do I install Efficient Embedded Browser Loop in Codex?
Run `npx skills add holaboss-ai/holaOS --skill browser-core-efficient -a codex`. Or copy the skill folder (runtime/harnesses/src/embedded-skills/browser-core-efficient in holaboss-ai/holaOS) into .agents/skills/browser-core-efficient in your project. Codex loads it when a task matches its description.
Can I use Efficient Embedded Browser Loop 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 holaboss-ai/holaOS --skill browser-core-efficient -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/browser-core-efficient, .gemini/skills/browser-core-efficient, .github/skills/browser-core-efficient and .opencode/skills/browser-core-efficient in your project.
What does Efficient Embedded Browser Loop need to run?
SKILL.md names no scripts, command-line tools or credentials: Efficient Embedded Browser Loop is instructions for the agent only. Our summary lists: The workspace's embedded browser tools, such as browser_act and browser_find.
Does Efficient Embedded Browser Loop access the network?
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Is Efficient Embedded Browser Loop 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 Efficient Embedded Browser Loop use?
Efficient Embedded Browser Loop has a licence file (the repository's licence) that doesn't match a standard licence. Read it on GitHub before reusing the skill.
How many tokens does Efficient Embedded Browser Loop use?
About 736 tokens (SKILL.md is roughly 2.9k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.
What are the alternatives to Efficient Embedded Browser Loop?
Skills that share tags, products or a category with Efficient Embedded Browser Loop: Agent Browser CLI (vercel-labs/agent-browser, 44k stars), Agent Browser (quran/quran.com-frontend-next, 1.9k stars), Web Access via Browser CDP (eze-is/web-access, 9.1k stars) and Dev Browser Automation (MemTensor/MemOS, 12k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
Who maintains Efficient Embedded Browser Loop?
holaboss-ai (a GitHub organization) maintains it in holaboss-ai/holaOS, which has 11,449 GitHub stars. The repository holds 30 skills in this directory. The repository was last updated on August 21, 2026.
Source: holaboss-ai/holaOS on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.