kimi-search_skill

This skill helps you perform real-time web searches using the Kimi API to fetch up-to-date information in Chinese and English.
  • Python

2.6k

GitHub Stars

2

Bundled Files

2 months ago

Catalog Refreshed

3 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill openclaw/skills --skill kimi-search

  • _meta.json276 B
  • SKILL.md4.8 KB

Overview

This skill integrates Moonshot Kimi's built-in $web_search tool to perform live web searches and return summarized answers. It requires a MOONSHOT_API_KEY and a Python environment with the openai package. Use it when you need real-time information or to augment the model with current web content.

How this skill works

The skill declares the $web_search tool to the Kimi model and forwards tool_call requests to the Kimi service so the search is executed server-side. The script returns the tool_call parameters and receives Kimi's search results, which it then formats into a JSON response containing the original query, a generated answer, and token usage. It uses the kimi-k2.5 model with temperature=0.6 and disables thinking mode to support tool calls.

When to use it

  • Retrieve breaking news, live events, or time-sensitive facts.
  • Verify current product details, pricing, or documentation updates.
  • Research topics where up-to-date web sources are required.
  • Augment agent responses with external citations or recent examples.
  • Run quick web lookups inside an OpenClaw agent or via the standalone script.

Best practices

  • Provision and protect MOONSHOT_API_KEY via environment variable for higher security.
  • Ensure account has sufficient balance; each search costs ~¥0.03 plus model token fees.
  • Use concise queries to reduce token usage; search results can consume 5k–10k tokens.
  • Restart the gateway after updating openclaw.json to apply API key changes.
  • Keep openai dependency installed in a venv to avoid system-wide conflicts.

Example use cases

  • Ask an agent to "search today's headlines" and receive a summarized update with sources.
  • Run the script: ~/.openclaw/skills/kimi-search/scripts/kimi-search "OpenClaw news" to get a JSON result.
  • Check the latest Python release notes or library changes before making code decisions.
  • Confirm live product availability or pricing on e-commerce sites when advising users.
  • Integrate into backup or archive workflows to tag records with current web references.

FAQ

The skill will fail with a missing key error. Configure the key in ~/.openclaw/openclaw.json or export MOONSHOT_API_KEY before running.

How much does each search cost?

Each web search incurs an extra ~¥0.03 fee in addition to model token charges listed on the Moonshot pricing page.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational