Skills Explorer Add Skills

kimi-use

An agent skill by daymade, from daymade/claude-code-skills. Tags: ai, automation, data, research, web.

What it does

Drive the Kimi desktop app (Kimi.app) through computer-use to query its built-in data plugins — 天眼查 company records, 同花顺 iFinD financials, 财新数据, 标普全球市场财智, 恒生聚源, SEC, IMF, 世界银行公开数据, 学术数据库, 法律数据库 and more — through the user's own logged-in Kimi session, with no separate API keys. Use whenever the user says "用 Kimi 查" / "操作 Kimi 客户端" / "Kimi 插件", asks to fetch company shareholders, financial statements, market data, or academic/legal records via Kimi, or needs a data source that has no standalone API but exists as a Kimi plugin. Covers both Claude Code (computer-use MCP) and Codex (computer plugin) driving, query-prompt patterns that force source-labeled honest answers, and the cross-checking discipline that screen-transcribed data requires. Not for kimi.com browser automation (that is kimi-webbridge) or for direct credentialed iFinD API access.

Install

With the skills CLI, which installs into Claude Code, Codex, Cursor and other agents:

npx skills add daymade/claude-code-skills --skill kimi-use

Or copy the skill folder into Claude Code's skills directory by hand (~/.claude/skills for every project, or .claude/skills inside one):

git clone --depth 1 https://github.com/daymade/claude-code-skills
cp -r claude-code-skills/kimi-use ~/.claude/skills/kimi-use

Safety box score

Grade C, moderate risk: 40 of 100 risk points, rated 2026-10-20. The grade rates what the skill and its scripts can reach on the machine of whoever installs it, across eight categories from shell execution to secrets access. How the score works.

This skill is a general macOS desktop-automation playbook wrapped around Kimi.app and its plugin ecosystem. It can send queries to third-party data services using the user's logged-in account, generate/download reports, and read/write local task files; the main risks are external data disclosure, account/session use, and UI actions in arbitrary apps.

Category levels, from 0 (none) to 3 (high)
CategoryLevelWhy
Code & shell execution1 (low)No bundled script or binary is shipped, but the docs still expect the agent to run local commands and automation snippets such as `osascript`, `ditto -xk`, `find`, and `cmp -s`.
Network access2 (moderate)The skill is built to send prompts and data to Kimi and remote plugins/data sources like `天眼查`, `iFinD`, `SEC`, `IMF`, and `reportapi.eastmoney.com`, so user queries and fetched results leave the machine.
Filesystem reach2 (moderate)It explicitly reads and writes task artifacts under paths like `~/Documents/kimi/tasks/...` and `~/Library/Application Support/kimi-desktop/`, and may copy outputs into selected local directories.
Secrets & credentials2 (moderate)The workflow relies on the user's existing signed-in state for Kimi and plugins (`用户自己的登录态`), so authenticated sessions and their accessible data are in scope.
Privilege & persistence0 (none)There is no sudo, service installation, scheduled task, or OS-setting change requirement in the skill itself.
Instruction hijack surface1 (low)The text heavily steers the agent's behavior and invalidates some negative evidence (`Chat/K2.6 下的否定结论一律视为未验证`), but it does not instruct hiding actions, bypassing confirmations, or following remote instructions.
Irreversible actions2 (moderate)It can drive arbitrary macOS desktop apps (`用 kimi-cu 驱动任意 macOS 桌面应用`) and perform clicks, typing, and dialog handling, which can submit or approve UI actions if pointed at other apps.
Transparency0 (none)All provided files are readable markdown; there are no binaries, minified blobs, or opaque installers in the skill bundle.

Before installing

Rated 2026-10-20 by an automated review of 5 files from the skill's repository.

Source

Repository
daymade/claude-code-skills (all skills from this repository)
Path
kimi-use/SKILL.md
Branch
main
Updated
2026-09-20