Update to Vercel

{
  "name": "vercel-api",
  "description": "Vercel app and REST API expert guidance. Use when the agent needs live access to Vercel projects, deployments, environment variables, domains, logs, or documentation through the connected Vercel app or REST API.",
  "included_files": [
    {
      "relative_path": "agents/openai.yaml",
      "size_in_bytes": 103
    }
  ],
  "skill_md_contents": "---\nname: vercel-api\ndescription: Vercel app and REST API expert guidance. Use when the agent needs live access to Vercel projects, deployments, environment variables, domains, logs, or documentation through the connected Vercel app or REST API.\nmetadata:\n  priority: 7\n  docs:\n    - \"https://vercel.com/docs/rest-api\"\n  sitemap: \"https://vercel.com/sitemap/docs.xml\"\n  pathPatterns:\n    - '.mcp.json'\n    - '.vercel/project.json'\n  bashPatterns:\n    - '\\bnpm\\s+(install|i|add)\\s+[^\\n]*@vercel/sdk\\b'\n    - '\\bpnpm\\s+(install|i|add)\\s+[^\\n]*@vercel/sdk\\b'\n    - '\\bbun\\s+(install|i|add)\\s+[^\\n]*@vercel/sdk\\b'\n    - '\\byarn\\s+add\\s+[^\\n]*@vercel/sdk\\b'\n    - '\\bmcp\\.vercel\\.com\\b'\n---\n\n# Vercel API — Connected App & REST API\n\nYou are an expert in the Vercel platform APIs. This plugin uses the connected Vercel app for live, authenticated access to Vercel resources, and this skill also covers the direct REST API when connector coverage is not enough.\n\n## Connected Vercel App\n\nThe plugin's `.app.json` points Codex at the connected Vercel app. The linked Vercel documentation still describes the underlying MCP server and REST API, and this skill uses that documentation where it helps explain capabilities and fallback paths.\n\n### Connection\n\n```\nURL:       https://mcp.vercel.com\nTransport: Streamable HTTP\nAuth:      OAuth 2.1 (automatic — agent is prompted to authorize on first use)\n```\n\nOn first connection the agent will open a browser-based OAuth flow to grant read access to your Vercel account. Subsequent sessions reuse the stored token.\n\n### Available MCP Tools\n\nThe connected Vercel app exposes these tool categories:\n\n| Category | Capabilities |\n|----------|-------------|\n| **Documentation** | Search and navigate Vercel docs, Next.js docs, AI SDK docs |\n| **Projects** | List projects, get project details, view project settings |\n| **Deployments** | List deployments, inspect deployment details, view build output |\n| **Logs** | Query deployment logs, function invocation logs, build logs |\n| **Domains** | List domains, check domain configuration and DNS status |\n| **Environment Variables** | List env vars per project and environment |\n| **Teams** | List teams, view team members and settings |\n\n### Usage Patterns\n\n#### Diagnose a failed deployment\n\n```\n1. List recent deployments → find the failed one\n2. Inspect deployment → get error summary\n3. Query build logs → identify root cause\n4. Cross-reference with vercel-functions skill for runtime fixes\n```\n\n#### Audit project configuration\n\n```\n1. Get project details → check framework, build settings, root directory\n2. List environment variables → verify required vars are set per environment\n3. List domains → confirm production domain is correctly assigned\n4. Check deployment logs → look for runtime warnings\n```\n\n#### Search documentation\n\n```\n1. Search Vercel docs for a topic → get relevant pages\n2. Read specific doc page → extract configuration examples\n3. Cross-reference with bundled skills for deeper guidance\n```\n\n#### Debug function performance\n\n```\n1. Query function logs → find slow invocations\n2. Inspect deployment → check function region, runtime, memory\n3. Cross-reference with vercel-functions skill for optimization patterns\n```\n\n## Deploying Your Own MCP Server\n\nUse the `mcp-handler` package (renamed from `@vercel/mcp-adapter`) to build and deploy custom MCP servers on Vercel with Next.js, Nuxt, or SvelteKit:\n\n```bash\nnpm install mcp-handler\n```\n\nMCP servers deployed on Vercel use **Streamable HTTP** transport (replaced SSE in March 2025 MCP spec) — cuts CPU usage vs SSE with no persistent connections required. Used in production by Zapier, Composio, Vapi, and Solana.\n\nSee [Deploy MCP servers to Vercel](https://vercel.com/docs/mcp/deploy-mcp-servers-to-vercel) and [GitHub: mcp-handler](https://github.com/vercel/mcp-handler).\n\n## REST API (Direct Access)\n\nWhen the MCP server doesn't cover a use case (or for write operations), use the Vercel REST API directly with `@vercel/sdk` or `curl`.\n\n### Authentication\n\n```bash\n# Bearer token auth (personal token or team token)\ncurl -H \"Authorization: Bearer $VERCEL_TOKEN\" https://api.vercel.com/v9/projects\n```\n\n```typescript\n// @vercel/sdk\nimport { Vercel } from '@vercel/sdk';\n\nconst vercel = new Vercel({ bearerToken: process.env.VERCEL_TOKEN });\n```\n\n### Key Endpoints\n\n| Endpoint | Method | Purpose |\n|----------|--------|---------|\n| `/v9/projects` | GET | List all projects |\n| `/v9/projects/:id` | GET | Get project details |\n| `/v13/deployments` | GET | List deployments |\n| `/v13/deployments` | POST | Create a deployment |\n| `/v13/deployments/:id` | GET | Get deployment details |\n| `/v9/projects/:id/env` | GET | List environment variables |\n| `/v9/projects/:id/env` | POST | Create environment variable |\n| `/v6/domains` | GET | List domains |\n| `/v6/domains` | POST | Add a domain |\n| `/v1/edge-config` | GET | List Edge Configs |\n| `/v1/firewall` | GET | List firewall rules |\n| `/v1/drains` | GET | List all drains |\n| `/v1/drains` | POST | Create a drain |\n| `/v1/drains/:id/test` | POST | Test a drain |\n| `/v1/drains/:id` | PATCH | Update a drain |\n| `/v1/drains/:id` | DELETE | Delete a drain |\n| `/v3/deployments/:id/events` | GET | Stream runtime logs |\n\n### SDK Examples\n\n#### List deployments\n\n```typescript\nimport { Vercel } from '@vercel/sdk';\n\nconst vercel = new Vercel({ bearerToken: process.env.VERCEL_TOKEN });\n\nconst { deployments } = await vercel.deployments.list({\n  projectId: 'prj_xxxxx',\n  limit: 10,\n});\n\nfor (const d of deployments) {\n  console.log(`${d.url} — ${d.state} — ${d.created}`);\n}\n```\n\n#### Manage environment variables\n\n```typescript\n// List env vars\nconst { envs } = await vercel.projects.getProjectEnv({\n  idOrName: 'my-project',\n});\n\n// Create env var\nawait vercel.projects.createProjectEnv({\n  idOrName: 'my-project',\n  requestBody: {\n    key: 'DATABASE_URL',\n    value: 'postgres://...',\n    target: ['production', 'preview'],\n    type: 'encrypted',\n  },\n});\n```\n\n#### Get project domains\n\n```typescript\nconst { domains } = await vercel.projects.getProjectDomains({\n  idOrName: 'my-project',\n});\n\nfor (const d of domains) {\n  console.log(`${d.name} — verified: ${d.verified}`);\n}\n```\n\n## Observability APIs\n\n### Drains (`/v1/drains`)\n\nDrains forward logs, traces, speed insights, and web analytics data to external endpoints. All drain management is REST API or Dashboard (`https://vercel.com/dashboard/{team}/~/settings/log-drains`) only — no CLI commands exist.\n\n```typescript\nimport { Vercel } from '@vercel/sdk';\n\nconst vercel = new Vercel({ bearerToken: process.env.VERCEL_TOKEN });\n\n// List all drains\nconst drains = await vercel.logDrains.getLogDrains({ teamId: 'team_xxxxx' });\n\n// Create a drain\nawait vercel.logDrains.createLogDrain({\n  teamId: 'team_xxxxx',\n  requestBody: {\n    url: 'https://your-endpoint.example.com/logs',\n    type: 'json',\n    sources: ['lambda', 'edge', 'static'],\n    environments: ['production'],\n  },\n});\n```\n\n> For payload schemas (JSON, NDJSON), signature verification, and vendor integration setup, see `⤳ skill: observability`.\n\n### Runtime Logs (`/v3/deployments/:id/events`)\n\nStream runtime logs for a deployment. The response uses `application/stream+json` — each line is a separate JSON object. Always set a timeout to avoid hanging on long-lived streams.\n\n```typescript\n// Query via MCP (recommended for agents)\n// Use the get_runtime_logs MCP tool for structured log access\n\n// Direct REST alternative (streaming)\nconst res = await fetch(\n  `https://api.vercel.com/v3/deployments/${deploymentId}/events`,\n  { headers: { Authorization: `Bearer ${process.env.VERCEL_TOKEN}` } }\n);\n// Parse as NDJSON — see observability skill for streaming code patterns\n```\n\n## `vercel api` CLI Command (January 2026)\n\nThe `vercel api` command gives agents direct access to the full Vercel REST API from the terminal with no additional configuration. It uses the CLI's existing authentication, so agents can call any endpoint immediately.\n\n```bash\n# Call any REST endpoint directly\nvercel api GET /v9/projects\nvercel api GET /v13/deployments\nvercel api POST /v9/projects/:id/env --body '{\"key\":\"MY_VAR\",\"value\":\"val\",\"target\":[\"production\"]}'\n```\n\nThis bridges the gap between the read-only MCP server and the full REST API — agents can perform write operations without needing `@vercel/sdk` or manual `curl` with tokens.\n\n## When to Use MCP vs CLI vs REST API\n\n| Scenario | Use | Why |\n|----------|-----|-----|\n| Agent needs to inspect/read Vercel state | **MCP server** | OAuth, structured tools, no token management |\n| Agent needs to deploy or mutate state | **CLI** (`vercel deploy`, `vercel env add`) | Full write access, well-tested |\n| Agent needs ad-hoc API access | **`vercel api`** | Direct REST from terminal, no token setup |\n| Programmatic access from app code | **REST API / @vercel/sdk** | TypeScript types, fine-grained control |\n| CI/CD pipeline automation | **CLI + VERCEL_TOKEN** | Scriptable, `--prebuilt` for speed |\n| Searching Vercel documentation | **MCP server** | Indexed docs, AI-optimized results |\n\n## Cross-References\n\n- **CLI operations** → `⤳ skill: vercel-cli`\n- **Function configuration** → `⤳ skill: vercel-functions`\n- **Storage APIs** → `⤳ skill: vercel-storage`\n- **Firewall rules** → `⤳ skill: vercel-firewall`\n- **AI SDK MCP client** → `⤳ skill: ai-sdk` (section: MCP Integration)\n- **Drains, log streaming, analytics export** → `⤳ skill: observability`\n\n## Official Documentation\n\n- [Vercel MCP](https://vercel.com/docs/mcp)\n- [Vercel REST API](https://vercel.com/docs/rest-api/reference)\n- [@vercel/sdk](https://www.npmjs.com/package/@vercel/sdk)\n- [MCP Authorization Spec](https://spec.modelcontextprotocol.io)\n- [GitHub: Vercel SDK](https://github.com/vercel/sdk)\n"
}