convex/convex-kimi-plugin

最近记录的来源活动: 2026年10月29日 16:51

SkillsMP 收录数据更新: 2026年11月9日 14:22

已收集 skills: 6

GitHub 星标: 1
GitHub Forks: 1

用 AI 助手安装

复制这段 Prompt，发给你正在使用的 AI 助手。

Follow https://skillsmp.com/skill-install/prompt.md to install Agent Skills from https://github.com/convex/convex-kimi-plugin.

仓库 README

仓库 README 来源：README.md

查看 README 源文件

展开完整 README

# convex-kimi-plugin The official [Convex](https://www.convex.dev) plugin for the [Kimi Code CLI](https://github.com/moonshotai/kimi-code). Adds a reactive, type-safe backend to JS/TS apps: scaffold a running Next.js + Convex app from one sentence (`$quickstart`), add capabilities from the Convex component ecosystem (`$add`), plus auth, billing, crons, domains, migrations, seeding, testing, a `convex-expert`, a `convex-reviewer`, and live error-watching via MCP. ## Install ``` /plugins install https://github.com/convex/convex-kimi-plugin ``` ## Layout - `plugins/marketplace.json` — plugin registry for this repo (generated). - `plugins/official/convex/kimi.plugin.json` — the plugin manifest (generated). - `plugins/official/convex/skills/<id>/SKILL.md` — one skill per capability (generated). - `plugins/official/convex/mcp/` — the `convex-plugin` MCP server (live error-watching). ## Generated — do not hand-edit The skills and manifests here are **generated** from the [convex-agents](https://github.com/convex/convex-agents) hub (`content/capabilities/*.json` → `generators/forge.mjs` + `generators/kimi.mjs`). Fix knowledge there, regenerate, and PR the result here. ## Privacy & data This plugin connects to Convex services and collects anonymous usage data. See the [Convex privacy policy](https://convex.dev/legal/privacy) for full details and your rights. Three kinds of data can leave your machine, each governed by a rule that holds no matter which command triggers it: ### 1. Anonymous usage telemetry (on by default, opt-out) Hooks may send anonymous telemetry to Convex's PostHog project: a random device id, the plugin version, your OS, and coarse event names (session start, lint/typecheck counts). Never your code, file paths, prompts, or personal identifiers. Opt out with `CONVEX_PLUGIN_TELEMETRY=0` or `DO_NOT_TRACK=1`. ### 2. Building your app (only when you invoke a scaffolding flow) Flows that scaffold or extend an app (such as `quickstart` and `/add`) send the inputs you give them to the Convex scaffolding service so it can build for you — for example, the one-sentence idea you type is sent to the scaffolding endpoint and logged as a run start. These flows also download and run setup scripts from that service. This happens only when you invoke such a flow. ### 3. Sharing a session to improve the tools (one-time, explicit opt-in) Some flows can offer to send a **redacted** copy of your current session to the Convex team to help improve these tools. This is **opt-in and asked once**: the first time it would send, you're asked to choose **Always**, **Just this once**, or **Never**. _Always_ and _Never_ are remembered per user (stored in `~/.convex/improve-consent`) so you're not asked again; _Just this once_ shares only that session and isn't stored, so a later session asks again. **Nothing is sent until you've explicitly chosen to share**: the flow is built to ask first, and the skill instructs the agent not to answer on your behalf. For a hard guarantee that nothing is ever sent, set `CONVEX_IMPROVE_CONSENT=never` — it's an absolute kill switch that overrides any stored choice. Secrets are redacted before anything leaves your machine on every send, regardless of your choice. Change your mind anytime by deleting `~/.convex/improve-consent` or setting that variable. If you don't invoke these flows, nothing beyond the anonymous telemetry above leaves your machine. ## License Apache-2.0

正在加载仓库 README…

这个仓库中的 skills

职业分类待补全

skill

职业分类

描述

更新

convex-expert

未分类

Convex backend rules — consult this whenever writing or editing any code inside a convex/ directory (schemas, queries, mutations, actions, HTTP endpoints, crons, file storage, auth, component installation). TRIGGER before touching convex/ functions, so the…

原文语言：英语

2026-10-29

quickstart

未分类

Scaffold a running Next.js + shadcn Convex app from one sentence — convex dev + next dev already running — then build the idea live (runs locally). TRIGGER when the user wants to START a new Convex app from scratch — e.g. they ran $quickstart, said 'scaffold…

原文语言：英语

2026-09-30

convex-authz

未分类

Audit and harden a Convex app's authorization: identity-from-arg impersonation, missing per-document ownership checks, public queries leaking PII/financial data by a client-supplied id, and writes into a parent/container the caller doesn't own — the single…

原文语言：英语

2026-08-28

check-updates

未分类

Check the CURRENT Convex app's pinned components against the latest recommended versions and offer to upgrade them — e.g. the passkey auth component's new email-first sign-in. TRIGGER when the user runs /check-updates or $check-updates, asks 'are my…

原文语言：英语

2026-07-22

convex-reviewer

未分类

Review Convex code for security, auth, validators, performance, and best practices. TRIGGER when the user asks to review/audit Convex code, or after writing convex/ functions you want checked. Applies the Convex-specific review checklist (auth checks,…

原文语言：英语

2026-07-22

domains

未分类

Wire a domain the user ALREADY OWNS (GoDaddy/Namecheap/Cloudflare/…) to their Convex app: exact DNS records, custom-domain attachment, auth-origin rebind. TRIGGER when the user owns a domain and wants it pointing at their app ('point my domain at this', 'use…

原文语言：英语

2026-07-22

已展示 6 / 6 个已收集 Skill。