← Back to Skills Marketplace
Use Browser (Scrape leads, like, post on socials or perform actions on the web)
by
mathiasthu
· GitHub ↗
· v1.0.0
583
Downloads
1
Stars
3
Active Installs
1
Versions
Install in OpenClaw
/install use-browser
Description
Automates browser interactions for social media management across Instagram, LinkedIn, and X. Handles posting, DMs, connection requests, lead scraping, and m...
Usage Guidance
Before installing, verify the source and installation method for the required 'browser-use' binary (the SKILL.md expects it but the registry metadata doesn't declare it). Inspect the GitHub repo and ensure you trust the binary provider and release artifacts. Confirm how and where session cookies/login state are stored on the VM, who can access them, and how long they persist — if possible, require ephemeral sessions or explicit user confirmation for sensitive actions. Consider restricting autonomous actions (require confirmation for sending messages/connection requests by default) and double-check the domain allowlist. If you plan to use production accounts, run this on a tightly isolated VM and avoid using high-value credentials until you understand the binary and storage behavior.
Capability Analysis
Type: OpenClaw Skill
Name: use-browser
Version: 1.0.0
The skill bundle is classified as suspicious due to the inclusion of powerful and high-risk commands like `browser-use:eval` and `browser-use:cookies` in SKILL.md. While the documentation provides extensive and explicit instructions for the AI agent to prevent misuse (e.g., strict domain allowlist, 'NEVER' rules for `eval` regarding network calls or cookie access, and robust prompt injection defenses), these are textual instructions. The inherent power of these commands, particularly `eval` which can execute arbitrary JavaScript within the browser context, presents a significant vulnerability if the agent's adherence to these instructions is compromised by a sophisticated prompt injection attack, potentially leading to data exfiltration or unauthorized actions despite the stated intent to prevent them.
Capability Assessment
Purpose & Capability
SKILL.md describes a browser-automation CLI ('browser-use') for Instagram/LinkedIn/X which is coherent with the declared purpose. However, the skill frontmatter (metadata.openclaw) requires the 'browser-use' binary while the registry metadata shows 'Required binaries: none' — a clear inconsistency. If the binary is genuinely required, it should be declared and an install mechanism provided or documented.
Instruction Scope
Instructions allow scraping profiles/leads and sending DMs/connection requests without confirmation and assume pre-authenticated, persistent sessions on an 'isolated VM'. The doc enforces a domain allowlist and blocks local/cloud-metadata ranges (good), but it does not state where scraped data, cookies, or session state are stored/transmitted, nor how long persistence lasts. Lack of guidance on data retention/exfiltration and the permission to act autonomously on messaging/connection requests are scope concerns.
Install Mechanism
There is no install spec (instruction-only), but the CLI 'browser-use' is required by the SKILL.md. The README links to a GitHub repo, but the skill provides no explicit install steps or vetted source for the binary. This gap makes it unclear who provides/maintains the executable and increases risk if the binary must be fetched manually.
Credentials
The skill declares no required environment variables or credentials, which aligns with the claim that sessions are pre-authenticated manually. However, persistent cookies/session storage are implied but not described (location, encryption, access controls), so sensitive authentication material may persist outside the user's control. No unrelated credentials are requested.
Persistence & Privilege
always is false and the skill is user-invocable, which is appropriate. However, because the agent is allowed to reuse persistent authenticated sessions and perform high-impact actions (send DMs, connection requests, posting, scraping) without confirmation, autonomous invocation could have a large blast radius. The skill lacks safer defaults (e.g., require confirmation for messaging by default), increasing risk.
How to Use
- Make sure OpenClaw is installed (local or Docker)
- Run the install command in chat:
/install use-browser - After installation, invoke the skill by name or use
/use-browser - Provide required inputs per the skill's parameter spec and get structured output
Version History
v1.0.0
browser-use 1.0.0 – Initial Release
- Automates browser-based social media management for Instagram, LinkedIn, and X, including posting, DMs, connection requests, lead scraping, and monitoring.
- Operates on an isolated VM with user-pre-authenticated sessions for persistent login and session continuity.
- Enforces domain allowlist; agent will only navigate to officially approved social platform domains.
- Provides high autonomy for typical tasks (posting, messaging, engagement), but pauses for user confirmation before any destructive or irreversible actions.
- Supports parallel sessions, browser switching (Safari default, Chrome fallback), and a versatile command set for navigation, interaction, and data extraction.
- Designed to be safe, configurable, and streamlined for social media workflows.
(effecitly scrape leads, like, post, etc on socials or perform actions on the web)
Metadata
Frequently Asked Questions
What is Use Browser (Scrape leads, like, post on socials or perform actions on the web)?
Automates browser interactions for social media management across Instagram, LinkedIn, and X. Handles posting, DMs, connection requests, lead scraping, and m... It is an AI Agent Skill for Claude Code / OpenClaw, with 583 downloads so far.
How do I install Use Browser (Scrape leads, like, post on socials or perform actions on the web)?
Run "/install use-browser" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.
Is Use Browser (Scrape leads, like, post on socials or perform actions on the web) free?
Yes, Use Browser (Scrape leads, like, post on socials or perform actions on the web) is completely free (open-source). You can download, install and use it at no cost.
Which platforms does Use Browser (Scrape leads, like, post on socials or perform actions on the web) support?
Use Browser (Scrape leads, like, post on socials or perform actions on the web) is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).
Who created Use Browser (Scrape leads, like, post on socials or perform actions on the web)?
It is built and maintained by mathiasthu (@mathiasthu); the current version is v1.0.0.
More Skills