Use Nextbrowser cloud API to spin up cloud browsers for Openclaw to run autonomous browser tasks. Primary use is creating browser sessions with profiles (per...
Security Analysis
medium confidenceThe skill's runtime instructions match its stated purpose (remote cloud browsers and automated browser tasks), but there are provenance and scope inconsistencies (undeclared config usage, no homepage/source, and explicit guidance to bypass approval for autonomous account actions) that raise concern.
The name/description (cloud browsers, profiles, autonomous browser tasks) align with the SKILL.md endpoints and examples (profiles, credentials, locations, chat/tasks). The API calls and features described are coherent with the stated purpose.
SKILL.md tells the agent to read the OpenClaw config key at skills.entries.next-browser.apiKey and to submit autonomous tasks that perform account actions (e.g., logging into Reddit accounts, upvoting, posting comments). The doc explicitly instructs always using skip_plan_approval=true (bypassing task-approval) and 'fast' mode — this reduces oversight and increases potential for automated misuse. The skill also instructs accessing stored credentials in Nextbrowser; these actions are within the service scope but are high-impact and require explicit user awareness and consent.
Instruction-only skill with no install script and no code files means nothing is written to disk by the skill itself. This is the lowest-risk install mechanism.
Registry metadata lists no required config paths or primary credential, but SKILL.md clearly requires an API key stored in OpenClaw config (skills.entries.next-browser.apiKey). That credential requirement is not declared in metadata. The only external secret needed is the Nextbrowser API key (reasonable for a cloud API) but the mismatch between declared requirements and runtime instructions is a configuration/provenance issue.
The skill is not force-installed (always:false). It allows model invocation (default), so the agent can call the skill autonomously. Combined with SKILL.md guidance to skip plan approvals for tasks, this raises the operational risk that the agent could perform automated account actions without interactive user confirmation. Autonomous invocation alone is normal; the explicit skip-plan guidance is what increases risk.
Guidance
This skill appears to do what it says (control cloud browsers via Nextbrowser) but has several red flags you should consider before installing: 1) Provenance unknown — there is no homepage or source repo and the registry owner is unverified; verify the vendor and review privacy/TOS before providing an API key. 2) Metadata mismatch — the SKILL.md expects an API key stored in OpenClaw config but the skill metadata does not declare that requirement; expect to store a Nextbrowser API key at skills.entries.next-browser.apiKey. 3) High-impact automation — the instructions explicitly encourage autonomous account actions (upvoting, posting comments) and recommend skipping approval; this can violate third-party platform policies and cause account bans or legal/ethical issues. 4) Least-privilege: if you proceed, use a dedicated Nextbrowser API key bound to an account with no valuable data, audit its activity, and avoid giving it access to real user accounts. 5) Request more info: ask the publisher for a homepage, privacy/security documentation, and an explanation why the registry metadata omits the config path. If you cannot verify the publisher or risks, do not install or only install in a tightly sandboxed environment.
Latest Release
v1.0.13
Version 1.0.13 - No file or documentation changes detected in this release. - No updates to features, endpoints, or usage instructions.
More by @highxshell
Published by @highxshell on ClawHub
