Online Search
MaliciousAudited by VirusTotal on May 10, 2026.
Overview
Type: OpenClaw Skill Name: online-search Version: 1.0.0 The skill provides a legitimate online search interface for the OpenClaw agent, utilizing a local Node.js script (prosearch.cjs) to communicate with a local proxy gateway (127.0.0.1). The code logic is straightforward, using structured JSON for requests to avoid injection vulnerabilities, and the instructions in SKILL.md are focused on ensuring search accuracy and preventing AI hallucinations. No evidence of data exfiltration, malicious execution, or harmful prompt injection was found.
Findings (0)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
Your search terms and filters are sent through the ProSearch gateway when the skill is used.
The script sends user-supplied search keywords and optional filters to a local ProSearch proxy endpoint. This is the skill's stated purpose, but it is still a networked tool action using user-provided query text.
const PROXY_HOST = '127.0.0.1'; const API_PATH = '/proxy/prosearch/search'; ... body.keyword = params.keyword; ... method: 'POST'
Use it when you intend to perform an online search, and avoid putting secrets or highly sensitive personal information into search queries.
Search activity may be associated with the logged-in service session handled by the local gateway.
The skill discloses that authentication is handled automatically by a background gateway using the user's logged-in state. This is coherent for the service, but it means searches may use delegated session/account context.
鉴权由后台网关自动处理（基于用户登录态），无需手动配置凭证。
Install only if you are comfortable with authenticated ProSearch/Yuanbao searches through the local gateway; use a different tool if you need anonymous searching.
Search snippets or links could be inaccurate, misleading, or contain prompt-like text that should not override your intent.
The provider's returned message, containing web result snippets and links, is inserted verbatim into the conversation and may then be used for analysis. That is expected for a search skill, but the retrieved content is still external and untrusted.
QClaw MUST output `message` verbatim as the primary search results ... QClaw may then add analysis or summary AFTER the verbatim results
Verify important claims from the linked sources and do not treat text from search results as instructions to the agent.