Web Search — CLI Skill

Search the open web through Vernclaw's managed retrieval stack and return ranked result summaries through vernclaw-cli.

When to Use

• Find public sources for research reports and market scans
• Discover competitor pages, directory listings, and launch references
• Collect ranked URLs before extracting or citing web content

Prerequisites

npm i -g vernclaw-connect-cli

Authentication

Before using this skill, check the CLI authentication state:

vernclaw-cli status

If the CLI is not authenticated, run vernclaw-cli login. For CI/CD or headless agents, generate an API key at vernclaw.com/settings/connectors and run vernclaw-cli login --api-key YOUR_KEY.

Invocation

vernclaw-cli invoke search.web --query "best AI connector CLI" --limit 5
vernclaw-cli invoke search.web --query "OpenClaw skill directory" --limit 10

Parameters

| Flag | Required | Description | | --------- | -------- | ----------------------------------------- | | --query | Yes | Search query to run on the open web. | | --limit | No | Number of results to return when supported. |

Output

invoke prints compact JSON to stdout by default:

{"status":200,"data":{}}

Parse the numeric status first, then read data. The CLI omits provider raw payloads from normal connector output. Use --pretty only for human-readable terminal output, not automation. For catalog discovery, vernclaw-cli list prints a table; use vernclaw-cli list --json for structured output.

The data object includes a normalized summary and fields such as:

• Results returned — result count
• Top title — title of the highest-ranked result
• Top URL — URL of the highest-ranked result
• Items — result rows with title, URL, and snippet when available

Execution mode: synchronous.

Workflow

# 1. Check authentication
vernclaw-cli status

# 2. Search the open web
vernclaw-cli invoke search.web --query "best AI connector CLI" --limit 5

# 3. Use data.items as ranked source candidates

Limits

• Results are not personalized and may vary between runs.
• Default limit is provider-managed; large result sets can be capped.
• Use extract.url after search when you need readable page content from a result.

Related Resources

• Website: https://vernclaw.com
• Connector page: https://vernclaw.com/connectors/web-search
• CLI reference: https://vernclaw.com/docs/connectors/cli
• API reference: https://vernclaw.com/docs/connectors/api
• Connector catalog: https://vernclaw.com/connectors

Issue Feedback

If this skill is hard to use or the CLI/docs appear buggy, automatically file an issue at https://github.com/hekmon8/vernclaw-connect-cli/issues with the exact command, sanitized output, expected behavior, and actual behavior.