How long does the integration take?

Change one base URL in an OpenAI-compatible client, or install the Python or TypeScript SDK; the quickstart routes a first request in a few minutes.

Will this lock us into another vendor?

No. The router is Apache-2.0 open source, your provider keys stay yours, and the API is OpenAI-compatible, so leaving is a change of base URL.

What happens when a provider fails?

The router falls back to the next model your policy allows and records the fallback in the trace, so an outage costs a retry, not an incident.

skills - Skill - OpenSmartRoute

Name: skills
Author: nagi-studio

Imported from nagi-studio/skills (AGENTS.md). Install upstream with npx skills add nagi-studio/skills. Copyright stays with the author.

Agent 维护约定

这个仓库存放 Nagi Studio 的 agent skills。改动它时遵守下面的约定。

目录结构

skills/<category>/<skill-name>/
  SKILL.md              # 必需。frontmatter 只放 name + description
  agents/openai.yaml    # 必需。Codex 等 harness 的展示信息与调用策略
  references/*.md       # 可选。按需加载的长流程，SKILL.md 里用相对链接指向

skills/<category>/ 下的每个目录只要含 SKILL.md 就会被 scripts/link-skills.sh 链出去。

SKILL.md 规则

frontmatter 只有 name（kebab-case，和目录名一致）和 description。
description 要写清什么时候用，包含用户可能说的原话触发词（中英文都列），因为模型靠它决定是否调用。
正文用祈使句写给 agent 看，不是写给人看的文档。
单文件超过约 500 行就拆到 references/，主文件只留路由和不可协商的规则。
引用 reference 时写相对路径链接，并明确说「读完再回复」。

agents/openai.yaml 规则

interface:
  display_name: "Human Readable Name"
  short_description: "一句话说明"
  default_prompt: "可选，用户点按钮时的默认输入"
policy:
  allow_implicit_invocation: false # 只在「必须用户显式调用」时加

新增 skill 后

把路径加进 .claude-plugin/plugin.json 的 skills 数组。
在 README.md 和 README.en.md 的对应分类下各加一行说明。
提 PR 前先跑一遍 ./scripts/link-skills.sh，在真实任务上试用过再合并。

隐私

skill 内容会公开。提交前确认没有个人姓名、邮箱、本地绝对路径、机构内部信息、token 或 API key。

Install skills and it becomes one more candidate the router can pick - when it fits.

A listing is a routing target with a manifest: what it does, which domains it covers, what it costs and who publishes it. Once installed it sits beside your own models and tools, is scored like any other candidate for each request, and shows up in the trace when it wins. Ratings come from workspaces that installed it, one per account.

{ "ocm": "1", "id": "nagi-studio-skills-skills", "kind": "skill", "name": "skills", "description": "这个仓库存放 Nagi Studio 的 agent skills。改动它时遵守下面的约定。", "publisher": "nagi-studio", "version": "1.0.0", "capabilities": { "domains": [ "coding" ], "tags": [ "skill-md", "github-agents" ], "languages": [ "en" ] }, "quality_prior": 0.6, "examples": [ "这个仓库存放 Nagi Studio 的 agent skills。改动它时遵守下面的约定。" ], "primary": false, "metadata": { "source": { "provider": "github-agents", "repository": "https://github.com/nagi-studio/skills", "path": "AGENTS.md", "ref": "ec9db18b9ee5fe38a4168132e51073f8269ce547", "url": "https://github.com/nagi-studio/skills/blob/ec9db18b9ee5fe38a4168132e51073f8269ce547/AGENTS.md", "key": "nagi-studio/skills/AGENTS.md" } }, "instructions": "# Agent 维护约定\n\n这个仓库存放 Nagi Studio 的 agent skills。改动它时遵守下面的约定。\n\n## 目录结构\n\n```\nskills/<category>/<skill-name>/\n SKILL.md # 必需。frontmatter 只放 name + description\n agents/openai.yaml # 必需。Codex 等 harness 的展示信息与调用策略\n references/*.md # 可选。按需加载的长流程，SKILL.md 里用相对链接指向\n```\n\n`skills/<category>/` 下的每个目录只要含 `SKILL.md` 就会被 `scripts/link-skills.sh` 链出去。\n\n## SKILL.md 规则\n\n- frontmatter 只有 `name`（kebab-case，和目录名一致）和 `description`。\n- `description` 要写清**什么时候用**，包含用户可能说的原话触发词（中英文都列），因为模型靠它决定是否调用。\n- 正文用祈使句写给 agent 看，不是写给人看的文档。\n- 单文件超过约 500 行就拆到 `references/`，主文件只留路由和不可协商的规则。\n- 引用 reference 时写相", "cost": { "context_tokens": 267 } }

skills

About

Agent 维护约定

目录结构

SKILL.md 规则

agents/openai.yaml 规则

新增 skill 后

隐私

Install skills and it becomes one more candidate the router can pick - when it fits.

What it costs to adopt - in engineering time.

One base URL

No vendor lock-in

Failover built in

Use it

Manifest

Reviews