smithery-mcp-deployment
Expert guidance for building, optimizing, and deploying Model Context Protocol (MCP) servers.
Install
mkdir -p .claude/skills/smithery-mcp-deployment && curl -L -o skill.zip "https://agentskills.codes/api/skills/download/414" && unzip -o skill.zip -d .claude/skills/smithery-mcp-deployment && rm skill.zipInstalls to .claude/skills/smithery-mcp-deployment
Activation
This is the description your AI agent reads to decide when to run this skill — the better it matches your request, the more reliably it fires.
Best practices for creating, optimizing, and deploying MCP servers to Smithery. Use this skill when:(1) Creating new MCP servers for Smithery deployment(2) Optimizing quality scores (achieving 90/100)(3) Troubleshooting deployment issues (0/0 tools, missing annotations, low scores)(4) Migrating existing MCP servers to Smithery format(5) Understanding Smithery's schema format requirements(6) Adding workflow prompts, tool annotations, or documentation resources(7) Configuring smithery.yaml and package.json for deploymentKey capabilities
- →Evaluates MCP schema compliance
- →Calculates deployment quality scores
- →Formats smithery.yaml configurations
- →Identifies missing tool annotations
- →Validates Zod/plain object schema compatibility
How it works
Analyzes repository structure against Smithery schema requirements and provides actionable feedback to reach specific score thresholds.
Inputs & outputs
When to use smithery-mcp-deployment
- →Optimize MCP server quality scores
- →Troubleshoot Smithery deployment errors
- →Configure smithery.yaml for deployment
- →Migrate MCP servers to Smithery format
About this skill
Smithery MCP Deployment Best Practices
Documentation Resources
Before implementing MCP features or troubleshooting issues, consult the official MCP specification:
Use the context7 tool to look up current MCP documentation:
- Primary resource:
https://context7.com/websites/modelcontextprotocol_io_specification - This provides the authoritative MCP specification for tools, prompts, resources, and protocol details
Critical: Schema Format (Most Common Issue)
The #1 cause of deployment failures is incorrect schema format. Smithery expects plain objects with Zod properties, NOT z.object() wrappers.
// WRONG - Results in "0/0 tools"
inputSchema: z.object({
param: z.string()
}).strict()
// CORRECT - Tools will be detected
inputSchema: {
param: z.string()
}
This applies to:
inputSchemain toolsoutputSchemain toolsargsSchemain prompts
Quality Scoring (90/100 Optimal)
| Feature | Points | How to Achieve |
|---|---|---|
| Tools with descriptions | 25 | Detailed 2-4 sentence descriptions |
| Tool annotations | 20 | Add inside config object (not 4th param) |
| Optional config | 15 | All fields optional or with defaults |
| Workflow prompts | 15 | Create 3-5 workflow prompts |
| Icon | 10 | Add icon.svg to repository root |
| Documentation | 5 | Comprehensive README |
Note: "Optional Config" (15pts) and "Config Schema" (10pts) are mutually exclusive. Optional is better UX and higher points.
Tool Registration Template
server.registerTool(
'tool_name',
{
title: 'Action-Oriented Title',
description: 'Clear 2-4 sentence description. Start with action verb. ' +
'Explain behavior (async/blocking). Mention related tools.',
inputSchema: {
param: z.string()
.describe('Specific description with examples (e.g., foo, bar)')
},
outputSchema: {
result: z.string()
},
annotations: { // Inside config object!
readOnlyHint: true, // Only reads data?
destructiveHint: false, // Deletes/destroys data?
idempotentHint: true, // Same input = same result?
openWorldHint: false // Deterministic results?
}
},
async (args) => {
// Args passed directly (not request.params.arguments)
return {
content: [{
type: 'text',
text: JSON.stringify(result, null, 2)
}]
};
}
);
Workflow Prompt Template
server.registerPrompt(
'workflow-name',
{
title: 'Workflow Title',
description: 'End-to-end workflow description',
argsSchema: { // Plain object, not z.object()!
// IMPORTANT: Only z.string() types supported
param: z.string().describe('Parameter description').optional()
}
},
async (args) => ({
messages: [{
role: 'user',
content: {
type: 'text',
text: `Multi-step workflow instructions...`
}
}]
})
);
Configuration Setup
src/types.ts:
export const configSchema = z.object({
apiToken: z.string().optional()
.describe("Token (or use API_TOKEN env var)"),
format: z.enum(["json", "markdown"]).default("markdown")
});
export type Config = z.infer<typeof configSchema>;
src/index.ts:
export { configSchema }; // Must export!
export default function createServer(config?: Config) {
const server = new McpServer({ name: 'your-server', version: '1.0.0' });
return server;
}
smithery.yaml (TypeScript runtime):
runtime: "typescript"
# Do NOT include configSchema - auto-detected from TypeScript export
Project Structure
your-mcp-server/
├── icon.svg <- REQUIRED for 10 points!
├── package.json
├── smithery.yaml
├── tsconfig.json
├── src/
│ ├── index.ts <- Export createServer & configSchema
│ ├── types.ts <- Define configSchema
│ ├── tools/ <- Tool implementations
│ ├── prompts/ <- Workflow prompts
│ └── resources/ <- Documentation resources
Testing Before Deployment
# 1. Lint TypeScript
npx tsc --noEmit
# 2. Build with Smithery
npm run build
# Look for: "Config schema: N fields (M required)"
# 3. Test with MCP Inspector
npx @modelcontextprotocol/inspector dist/index.js
# 4. Verify in Inspector:
# - tools/list shows all tools with annotations
# - prompts/list shows all prompts
# - Try calling each tool
Common Issues Quick Reference
| Symptom | Cause | Fix |
|---|---|---|
| "0/0 tools" | Using z.object() | Use plain objects for schemas |
| "9/16 parameters" | .optional()/.default() on schema | Remove modifiers, handle in handler |
| No annotations | Wrong placement | Put inside config object, not 4th param |
| Score stuck at 43 | Schema + annotations | Fix both issues |
| Icon not showing | Wrong location | Place icon.svg in repo root |
TS error: request.params | Old handler pattern | Use async (args) => not async (request) => |
| TS error: prompt argsSchema | Non-string types | Use only z.string().optional() |
| Deployment fails | configSchema in yaml | Remove from yaml for TypeScript runtime |
Detailed Documentation
For comprehensive guides, see:
- Schema Format Details: references/schema-format.md
- Quality Scoring Guide: references/quality-scoring.md
- Troubleshooting: references/troubleshooting.md
- Complete Examples: references/examples.md
- Migration Guide: references/migration.md
Path to 90/100 Checklist
- Use plain object schemas (not
z.object()) - Remove
.optional()/.default()from schemas (handle in handler) - Add comprehensive tool descriptions (2-4 sentences)
- Include annotations in all tools (inside config object)
- Create 3-5 workflow prompts (use
z.string().optional()for args) - Add icon.svg to repository root
- Make all config optional or with defaults
- Export configSchema from index.ts
- Remove configSchema from smithery.yaml (for TypeScript)
- Lint with
npx tsc --noEmitbefore building - Test locally with MCP Inspector
When not to use it
- →When deploying non-MCP server architectures
- →When writing standard client-side applications
Limitations
- →Scores are estimates based on provided rubric
- →Cannot guarantee environment-specific deployment success
How it compares
It performs automated audit-based improvements based on specific platform deployment scores rather than general code review.
Compared to similar skills
smithery-mcp-deployment side by side with the closest alternatives in the catalog.
| Skill | Installs | Updated | Safety | Difficulty |
|---|---|---|---|---|
| smithery-mcp-deployment (this skill) | 8 | 8mo | Review | Intermediate |
| apollo-deploy-integration | 1 | 27d | Caution | Intermediate |
| gcp-cloud-run | 5 | 5mo | Review | Intermediate |
| flow-nexus-platform | 6 | 4mo | Review | Beginner |
Try saying
Example prompts that trigger this skill in your AI assistant.
You might also like
apollo-deploy-integration
jeremylongshore
Deploy Apollo.io integrations to production. Use when deploying Apollo integrations, configuring production environments, or setting up deployment pipelines. Trigger with phrases like "deploy apollo", "apollo production deploy", "apollo deployment pipeline", "apollo to production".
gcp-cloud-run
aj-geddes
Deploy containerized applications on Google Cloud Run with automatic scaling, traffic management, and service mesh integration. Use for container-based serverless computing.
flow-nexus-platform
ruvnet
Comprehensive Flow Nexus platform management - authentication, sandboxes, app deployment, payments, and challenges
deployment-pipeline-design
wshobson
Design multi-stage CI/CD pipelines with approval gates, security checks, and deployment orchestration. Use when architecting deployment workflows, setting up continuous delivery, or implementing GitOps practices.
vercel-deployment
davila7
Expert knowledge for deploying to Vercel with Next.js Use when: vercel, deploy, deployment, hosting, production.
netlify-deploy
openai
Deploy web projects to Netlify using the Netlify CLI (`npx netlify`). Use when the user asks to deploy, host, publish, or link a site/repo on Netlify, including preview and production deploys.