tavily-hikari-research

Run multi-step Tavily research through Tavily Hikari with tvly-hikari.

安装方式

默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。

用 Codex 或 Claude 帮你安装复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。

直接命令不会经过审查 Prompt；运行前请先检查来源。

npx skills add https://github.com/IvanLi-CN/tavily-hikari --skill tavily-hikari-research

命令会保持在同一行。复制前请横向滚动并检查完整内容。

想先保存到本地？可下载 SkillsMP 当前能够提供的文件。

决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。

SKILL.md

来源说明 · 只读预览

Guide agents to use Tavily Hikari safely through tvly-hikari.

Run Tavily CLI workflows through a configured Tavily Hikari deployment.

Crawl bounded sites through Tavily Hikari with tvly-hikari.

name	tavily-hikari-research
description	Run multi-step Tavily research through Tavily Hikari with tvly-hikari.

Use this skill when the task needs multi-step Tavily research through Tavily Hikari.

Write a concise research question with scope, date, geography, and source constraints.

Run:

tvly-hikari research "Compare Tavily CLI, Agent Skills, and MCP for AI agents" --json

Save the response for longer investigations:

tvly-hikari research "Compare Tavily CLI, Agent Skills, and MCP for AI agents" --json -o tavily-research.json

If the official CLI returns an async request id, inspect or wait for the result through the official follow-up commands exposed by the wrapper:
```
tvly-hikari research status <request-id> --json
tvly-hikari research poll <request-id> --json -o tavily-research.json
```

Requests go to the configured Hikari /api/tavily/research facade.
Hikari token quota applies to research calls.
Hikari audit logs retain the downstream token context while upstream keys remain private.