Agent skill
build-mcp-use-server
Quellcode ansehen: yigitkonur/skills-by-yigitkonur
Installation
npx skills add yigitkonur/skills-by-yigitkonur --skill build-mcp-use-server39
Installationen
EU-hosted inference API
Power your AI agent skills with open-source models.
Drop-in OpenAI-compatible API. No data leaves Europe.
MiniMax
MiniMax M3
$0.40 / $1.40
per M tokens
Z.ai
GLM 5.3 Flash
$0.20 / $0.60
per M tokens
MoonshotAI
Kimi K3
$4.00 / $18.00
per M tokens
DeepSeek
DeepSeek V4.1 Flash
$0.40 / $1.40
per M tokens
Build mcp-use Server
Server-side mechanics for mcp-use v2 TypeScript MCP servers (the beta npm dist-tag line). This skill owns the v2 API surface; sister skills own structure, clients, agents, and the raw SDK. v1 (mcp-use/server, npm latest) appears here only as migration material.
When to use this skill
Trigger when the target code or request involves any of these:
- Importing
MCPServerfrommcp-use(root), or code still importingmcp-use/serverthat must move to v2. - Defining tools (definition-first
{ name, description, inputSchema, outputSchema }+ callback), resources, or prompts with zod v4 / Standard Schema. - Server-side
ctxwork —ctx.auth,ctx.sendLog,ctx.reportProgress,ctx.sendNotification,ctx.client.capabilities(),ctx.requestState, elicitation re-entry. - OAuth via
mcp-use/oauth/*providers (clerk, auth0, workos, supabase, keycloak, better-auth) oroauthCustomProvider. - MCP Apps / ChatGPT Apps views —
views/<name>/view.tsx, the toolviewfield,structuredContentprops, CSP metadata,mcp-use/reacthooks (useToolContext,useCallTool,useViewState). - Streamable HTTP serving —
server.listen,server.fetch,toNodeHandler(mcp-use/node),withMcpUse/createNextHandler(mcp-use/next). - Running
mcp-use dev | build | typecheck | start | deploy, the Inspector, tunneling, or the curl handshake on/mcp. - Production hardening and deploys: mcp-use Cloud, Vercel, Cloudflare Workers, Google Cloud Run, Supabase, Deno, Bun, Hono, Railway, Docker.
- Migrating from
mcp-usev1, raw@modelcontextprotocol/sdkservers, or OpenAI Apps-SDK widgets.
Do not use this skill when:
- The code is app-side
MCPClient/ browser / react client mounting — route tobuild-mcp-use-client. - The work is
MCPAgentLLM orchestration over MCP tools — route tobuild-mcp-use-agent. - The user wants raw official SDK primitives or strict stdio without mcp-use — route v1
@modelcontextprotocol/sdkwork tobuild-mcp-server-sdk-v1, and split-package v2@modelcontextprotocol/{core,server,client}work tobuild-mcp-server-sdk-v2(mcp-use v2 cannot serve stdio). - The question is layer placement, import direction, or composition-root structure — route to
build-clean-mcp-architecturefirst (seereferences/00-clean-architecture-coordination.md), then return here for mechanics.
Version stance
This skill teaches v2 (verified against mcp-use@2.0.0-beta.66; exact pins and drift policy in references/00-version-drift.md). Detect which world the project is in before applying anything:
from "mcp-use/server"anywhere → v1 project → start atreferences/28-migration/02-v1-to-v2-overview.md.- Root
MCPServerimport,views/,mcp-use/oauth/*→ v2 project → apply this skill directly. - Bare
npm install mcp-useinstalls v1 (latest); v2 needs thebetatag.
Coordinate with neighboring skills
| Skill | Owns | Handoff |
|---|---|---|
build-clean-mcp-architecture |
Folder layout, import direction, layer boundaries, composition root. | Read first for placement; this skill second for APIs. See references/00-clean-architecture-coordination.md. |
build-mcp-use-client |
MCPClient, browser/react client mounting, code mode. |
Hand off when the code stops being the server. |
build-mcp-use-agent |
MCPAgent orchestration where an LLM picks tools. |
Hand off when the work is the agent loop. |
build-mcp-server-sdk-v1 / build-mcp-server-sdk-v2 |
Raw official SDK servers, stdio-only constraints. | Hand off if the user forbids mcp-use or needs stdio. |
convert-mcp-sdk-v1-to-v2 |
Raw-SDK v1→v2 conversions. | Hand off for raw-SDK migrations; this skill covers mcp-use migrations. |
test-by-mcpc-cli |
Live mcpc session verification once a server runs. |
Use after this skill produces a running server. |
Detect intent
| Intent | Start here | Then read |
|---|---|---|
| Extend an existing v2 server | scripts/audit-server-readiness.sh.md |
references/04-tools/01-overview.md, references/05-responses/01-overview-decision-table.md, references/08-server-config/01-mcp-server-constructor.md, references/22-validate/01-inspector-walkthrough.md |
| v1 project detected | references/28-migration/02-v1-to-v2-overview.md |
references/28-migration/03-v1-to-v2-imports-server-and-tools.md, references/28-migration/04-v1-to-v2-responses-and-helpers.md, then per-area migration files |
| Greenfield HTTP tool server | references/02-setup/02-scaffold-with-create-mcp-use-app.md or references/02-setup/04-manual-http-server.md |
references/04-tools/02-registering-a-tool.md, references/05-responses/01-overview-decision-table.md, references/22-validate/02-curl-handshake.md, references/30-workflows/01-greenfield-tool-server-to-vercel.md |
| MCP Apps view / ChatGPT app | references/18-mcp-apps/canonical-anchor.md |
references/18-mcp-apps/server-surface/01-tool-view-field.md, references/18-mcp-apps/view-react/01-setup-and-providers.md; if the host/model must act on mounted UI, references/18-mcp-apps/view-react/09-useviewtool.md; then references/30-workflows/02-views-app-chart-widget.md, references/20-inspector/08-debugging-chatgpt-apps.md |
| Next.js drop-in | references/19-nextjs-drop-in/01-overview-withmcpuse.md |
references/19-nextjs-drop-in/02-route-and-file-placement.md, references/30-workflows/05-nextjs-drop-in.md |
| Auth / OAuth | references/11-auth/01-overview.md |
references/11-auth/02-attaching-a-provider.md, the provider file under references/11-auth/providers/, references/11-auth/06-debugging-checklist.md |
| Wrap a REST API | references/17-advanced/03-openapi-fromopenapi.md |
references/30-workflows/06-openapi-to-mcp.md |
| Compose/proxy other MCP servers | references/17-advanced/01-proxy-and-gateway.md |
references/17-advanced/02-proxy-auth-and-namespacing.md, references/30-workflows/07-proxy-gateway.md |
| State, notifications, elicitation | references/01-concepts/04-stateless-model-and-request-state.md |
references/10-sessions/03-state-patterns-without-sessions.md, references/14-notifications/01-overview.md, references/12-elicitation/01-overview.md, references/30-workflows/08-elicitation-input-required-flow.md |
| Deploy or production hardening | references/25-deploy/01-decision-matrix.md |
references/25-deploy/02-pre-deploy-checklist.md, the platform file under references/25-deploy/platforms/, references/24-production/04-security-hardening.md |
| Troubleshoot a concrete error | references/00-symptom-index.md |
references/27-troubleshooting/02-quick-diagnostic-table.md, references/27-troubleshooting/06-decision-tree.md |
Use references/00-reference-index.md only when the intent table is not specific enough and you need an exact filename.
Core rules
- Import
MCPServerand server APIs frommcp-use(root).mcp-use/serverdoes not exist in v2. - Install the
betatag:mcp-use@beta,@mcp-use/cli@beta, scaffold withcreate-mcp-use-app@beta. Require Node >= 22.22.2 and ESM; install aStandardSchemaWithJSONlibrary in the project (this skill's examples use zod v4). - Return raw MCP result envelopes (
CallToolResultetc.). The v1 helpers still exported are deprecated — onlyreferences/05-responses/07-deprecated-v1-helpers.mdteaches them, for migration. - v2 is stateless per request: no session stores in the shipped beta, no post-response push, no
ctx.sample(). Cross-request state goes through therequestStatecodec or your own store. - Serve over Streamable HTTP only (
/mcpby default). Strict stdio is a raw-SDK requirement — route out. - For views:
views/<name>/view.tsx, tool-levelviewfield,outputSchemarequired, props viastructuredContent, hooks frommcp-use/react. - Work in the actual package or subdirectory the user named; prefer improving an existing server over replacing it.
- Never claim the server is scaffolded, runnable, or verified in a read-only or plan-only run.
- For version-sensitive claims, read
references/00-version-drift.mdbefore editing examples or migration guidance.
Workflow
1. Lock target path and execution mode
Identify the concrete path to inspect and edit. Treat the run as plan-only when the environment is read-only, installs are blocked, or the user asked for analysis. Plan-only output must include exact files, install commands, implementation steps, and validation commands — and must not claim runtime validation.
2. Scan what already exists
Inspect the target for:
package.json:mcp-useversion (1.x = v1, 2.x = v2),zodmajor,"type": "module",@mcp-use/cli- imports:
mcp-useroot vsmcp-use/server(v1 signal),mcp-use/oauth/*,mcp-use/react,mcp-use/next - server shape:
new MCPServer(...), tool definitions,views/directory,view:fields,oauthconfig - runtime signals:
.mcp-use/build/, Docker, platform config, custom routes, middleware
For existing servers run scripts/audit-server-readiness.sh (usage: scripts/audit-server-readiness.sh.md). For version drift run scripts/check-mcp-use-version.sh (usage: scripts/check-mcp-use-version.sh.md).
Summarize: target path, v1 vs v2, existing server vs none, tools-only vs views, implementation-capable vs plan-only, chosen entry file.
3. Choose the branch
v1 project: do not extend v1 — migrate first via references/28-migration/02-v1-to-v2-overview.md unless the user explicitly wants v1 maintenance (then say the skill's guidance is v2 and scope the change narrowly).
Existing v2 server: follow the intent row for the requested change, then audit nearby mechanics (tools/schemas, results, config, auth, views, deploy).
No server: scaffold with create-mcp-use-app@beta (scripts/scaffold-mcp-use-server.sh automates it) or hand-build from references/02-setup/04-manual-http-server.md. For an existing app, add a side-car per references/02-setup/05-add-to-existing-app.md; for Next.js follow references/19-nextjs-drop-in/.
Underspecified: infer from the existing project when possible; ask only for user-owned choices that block implementation (the exposed service/data, auth policy, tools-only vs. views, or deploy target).
4. Build or extend
Default sequence:
- entry file and runtime shape (
references/02-setup/,references/09-transports/04-runtime-adapters-node-next-fetch.md) MCPServerconfig (references/08-server-config/)- tools with precise zod v4 schemas (
references/04-tools/) - resources/prompts only when they improve the interface (
references/06-resources/,references/07-prompts/) - auth, notifications, elicitation, views, proxy — only when the intent requires them
- custom health routes, logging, security hardening when shipping beyond local dev (
references/24-production/)
5. Validate
Pick the smallest set that proves the changed behavior; report only the rung reached.
- read-only scan: files inspected, nothing executed
- static:
mcp-use typecheck/npm run build - local runtime:
mcp-use devserving/mcp; its CLI-owned listener auto-mounts Inspector at/mcp/inspectorunless--no-inspectoris set. Directserver.listen(),server.fetch, embedded Next.js handlers, and plainmcp-use startdo not auto-mount it. - Inspector: surface observed and called (
references/22-validate/01-inspector-walkthrough.md) - curl protocol probe: use the native modern wire (no
initialize) fromreferences/09-transports/02-streamable-http.md, or the legacy compatibility handshake (initialize,tools/list,tools/call) inreferences/22-validate/02-curl-handshake.md; legacy POSTs requireAccept: application/json, text/event-stream. - unit:
server.fetch(new Request(...))tests (references/22-validate/04-unit-testing-server-fetch.md) - live client:
mcp-use client, tunnel, ortest-by-mcpc-clisession (references/22-validate/03-connect-real-clients.md) - deployed: health route plus a live MCP call against the public URL
For views, verify the text fallback (content) and, when possible, Inspector CSP mode (references/23-debug/03-view-debugging.md). For deploys, walk references/25-deploy/02-pre-deploy-checklist.md.
Decision rules
- Return concise complete
contentalways; addstructuredContentwhen there is anoutputSchema, a typed consumer, or view props — and keep the two semantically equivalent. - Put private or bulky data in
_meta; treatstructuredContentas model-visible (references/05-responses/06-meta-and-private-data.md). - Expected failures return
isErrorenvelopes; unexpected failures throw (references/05-responses/05-error-handling.md). - Guard the exact elicitation mode before returning
input_required: checkctx.client.capabilities().elicitation?.formforinputRequired.elicit(...)and.urlforinputRequired.elicitUrl(...)(references/12-elicitation/01-overview.md,references/16-client-introspection/02-capabilities.md). - Need model-side generation? The host generates, the tool validates — sampling is gone (
references/13-sampling/01-sampling-removed-in-v2.md). - Views:
view.namemust match the one-levelviews/<name>/view.tsxfolder;outputSchemais mandatory; declare domains inview.csp; useuseCallToolfor View → server anduseViewToolfor host/model → mounted View, never rawfetch. - One server definition serves both MCP Apps and ChatGPT hosts — never hand-roll
window.openai(references/18-mcp-apps/chatgpt-apps/01-dual-protocol.md).
Guardrails
- In an mcp-use server, do not mix in raw official SDK server primitives (
@modelcontextprotocol/sdkv1 or@modelcontextprotocol/{core,server}v2); route a raw-SDK implementation to the matching sibling skill instead. - Never install bare
mcp-usefor a v2 project — that is v1; pin thebetatag. - Never use zod v3, CommonJS, or Node < 22 with v2.
- Never use
z.any()/z.unknown()where a concrete schema is possible;.describe()every model-filled field. - Never teach or write v1 response helpers in new code; raw envelopes only.
- Never put secrets in source, logs,
structuredContent, view props, or view state. - Decide Host validation, Origin validation, and CORS separately: localhost
listen()auto-enables Host checks;server.fetch()needs explicitallowedHosts; Origin checks requireallowedOrigins; CORS headers requirecors(references/08-server-config/03-cors-and-allowed-origins.md,references/08-server-config/04-dns-rebinding-and-host-validation.md). - Never embed the MCP server as middleware inside another framework's app — use the runtime adapters or side-car patterns.
- Never claim session persistence exists in shipped v2 — see
references/10-sessions/01-overview-stateless-truth.md.
Validate honestly
| Rung | Evidence |
|---|---|
| Read-only scan | Files and references inspected; nothing ran. |
| Static validation | mcp-use typecheck / build passed. |
| Local runtime | mcp-use dev/start served /mcp. |
| Inspector | Connected; surface observed or called. |
| curl handshake | initialize, tools/list, and a relevant tools/call succeeded. |
| Live client | mcp-use client / mcpc session named with commands run. |
| Deployed endpoint | Health + live MCP operation verified on the public URL. |
For plan-only runs, mark runtime validation blocked and provide the exact commands to run later.
Output contract
Unless the user asks for another format, report:
- target path, v1-or-v2 verdict, and scan summary
- chosen branch and entrypoint decision
- implementation or exact plan
- validation rung reached, commands run, and blockers
- if views changed: text fallback and CSP verification state
- if deploy/production changed: platform deployment ID, source branch/SHA when available, terminal conclusion, exact dashboard/platform URL, checklist state, and live health/operation evidence
- key references used, with exact paths for the route actually followed
Reference routing
Start with intent or symptoms; use the inventory only as fallback.
- Symptom index:
references/00-symptom-index.md - Version drift:
references/00-version-drift.md - Clean architecture handoff:
references/00-clean-architecture-coordination.md - Full inventory:
references/00-reference-index.md - Bundled scripts:
scripts/check-mcp-use-version.sh.md,scripts/audit-server-readiness.sh.md,scripts/scaffold-mcp-use-server.sh.md - Concepts:
references/01-concepts/01-what-is-mcp-use.md,references/01-concepts/02-server-vs-client-vs-agent.md,references/01-concepts/03-transports-overview.md,references/01-concepts/04-stateless-model-and-request-state.md,references/01-concepts/05-mcp-spec-version-history.md,references/01-concepts/06-mcp-apps-and-views-terminology.md,references/01-concepts/07-this-skill-vs-build-mcp-use-client.md - Setup:
references/02-setup/01-prerequisites.md,references/02-setup/02-scaffold-with-create-mcp-use-app.md,references/02-setup/03-template-flags.md,references/02-setup/04-manual-http-server.md,references/02-setup/05-add-to-existing-app.md,references/02-setup/06-package-scripts.md,references/02-setup/07-tsconfig-and-types.md,references/02-setup/08-env-vars.md - CLI:
references/03-cli/01-overview.md,references/03-cli/02-create-mcp-use-app.md,references/03-cli/03-mcp-use-dev.md,references/03-cli/04-mcp-use-build-and-typecheck.md,references/03-cli/05-mcp-use-start.md,references/03-cli/06-mcp-use-deploy-and-cloud.md,references/03-cli/07-login-and-org.md,references/03-cli/08-client-and-screenshot.md,references/03-cli/09-flag-reference.md,references/03-cli/10-environment-variables.md - Tools:
references/04-tools/01-overview.md,references/04-tools/02-registering-a-tool.md,references/04-tools/03-schemas-standard-schema-and-zod-v4.md,references/04-tools/04-describe-and-annotations.md,references/04-tools/05-the-ctx-object.md,references/04-tools/06-validation-pipeline.md,references/04-tools/07-input-schema-vs-output-schema.md,references/04-tools/08-tool-anti-patterns.md,references/04-tools/canonical-anchor.md - Results:
references/05-responses/01-overview-decision-table.md,references/05-responses/02-text-and-content-blocks.md,references/05-responses/03-structured-content-and-output-schema.md,references/05-responses/04-images-audio-binary-resources.md,references/05-responses/05-error-handling.md,references/05-responses/06-meta-and-private-data.md,references/05-responses/07-deprecated-v1-helpers.md,references/05-responses/canonical-anchor.md - Resources:
references/06-resources/01-overview.md,references/06-resources/02-static-resources.md,references/06-resources/03-resource-templates.md,references/06-resources/04-binary-and-image.md,references/06-resources/05-uri-conventions.md,references/06-resources/06-subscriptions-listen.md,references/06-resources/canonical-anchor.md - Prompts:
references/07-prompts/01-overview.md,references/07-prompts/02-static-prompts.md,references/07-prompts/03-prompt-templates.md,references/07-prompts/04-completable-arguments.md,references/07-prompts/05-prompt-engineering.md - Server config:
references/08-server-config/01-mcp-server-constructor.md,references/08-server-config/02-network-basepath-and-endpoints.md,references/08-server-config/03-cors-and-allowed-origins.md,references/08-server-config/04-dns-rebinding-and-host-validation.md,references/08-server-config/05-middleware.md,references/08-server-config/06-custom-routes.md,references/08-server-config/07-lifecycle-listen-fetch-shutdown.md - Transports:
references/09-transports/01-overview.md,references/09-transports/02-streamable-http.md,references/09-transports/03-stateless-and-request-state.md,references/09-transports/04-runtime-adapters-node-next-fetch.md,references/09-transports/05-no-stdio-and-sse-history.md - Sessions/state:
references/10-sessions/01-overview-stateless-truth.md,references/10-sessions/02-session-storage-roadmap.md,references/10-sessions/03-state-patterns-without-sessions.md,references/10-sessions/04-multi-instance-and-scaling.md - Auth:
references/11-auth/01-overview.md,references/11-auth/02-attaching-a-provider.md,references/11-auth/03-ctx-auth-and-user-context.md,references/11-auth/04-permission-guards.md,references/11-auth/05-custom-provider-oauthcustomprovider.md,references/11-auth/06-debugging-checklist.md,references/11-auth/07-oauth-proxy-removed.md,references/11-auth/providers/01-clerk.md,references/11-auth/providers/02-auth0.md,references/11-auth/providers/03-workos.md,references/11-auth/providers/04-supabase.md,references/11-auth/providers/05-keycloak.md,references/11-auth/providers/06-better-auth.md - Elicitation:
references/12-elicitation/01-overview.md,references/12-elicitation/02-form-mode.md,references/12-elicitation/03-url-mode.md,references/12-elicitation/04-multi-round-and-request-state.md,references/12-elicitation/05-anti-patterns.md - Sampling:
references/13-sampling/01-sampling-removed-in-v2.md - Notifications:
references/14-notifications/01-overview.md,references/14-notifications/02-ctx-sendnotification.md,references/14-notifications/03-progress-reporting.md,references/14-notifications/04-list-changed-events.md,references/14-notifications/05-subscriptions-delivery.md,references/14-notifications/canonical-anchor.md - Logging:
references/15-logging/01-overview.md,references/15-logging/02-ctx-sendlog.md,references/15-logging/03-server-and-request-logging.md - Client introspection:
references/16-client-introspection/01-overview.md,references/16-client-introspection/02-capabilities.md,references/16-client-introspection/03-apps-detection.md,references/16-client-introspection/canonical-anchor.md - Advanced:
references/17-advanced/01-proxy-and-gateway.md,references/17-advanced/02-proxy-auth-and-namespacing.md,references/17-advanced/03-openapi-fromopenapi.md,references/17-advanced/04-mcp-use-vs-official-sdk.md,references/17-advanced/canonical-anchor.md - MCP Apps:
references/18-mcp-apps/01-what-are-mcp-apps.md,references/18-mcp-apps/02-mcp-apps-vs-chatgpt-apps-sdk.md,references/18-mcp-apps/03-vocabulary-views.md,references/18-mcp-apps/04-when-to-use-vs-tools-only.md,references/18-mcp-apps/05-host-capability-detection.md,references/18-mcp-apps/anti-patterns.md,references/18-mcp-apps/canonical-anchor.md - MCP Apps server surface:
references/18-mcp-apps/server-surface/01-tool-view-field.md,references/18-mcp-apps/server-surface/02-register-views-and-folder-conventions.md,references/18-mcp-apps/server-surface/03-viewconfig.md,references/18-mcp-apps/server-surface/04-assets-mcp-url-and-serving.md,references/18-mcp-apps/server-surface/05-csp-metadata.md - MCP Apps react:
references/18-mcp-apps/view-react/01-setup-and-providers.md,references/18-mcp-apps/view-react/02-usetoolcontext.md,references/18-mcp-apps/view-react/03-usecalltool.md,references/18-mcp-apps/view-react/04-useviewstate-and-model-context.md,references/18-mcp-apps/view-react/05-display-modes.md,references/18-mcp-apps/view-react/06-followups-and-open-external.md,references/18-mcp-apps/view-react/07-host-context-files-and-size.md,references/18-mcp-apps/view-react/08-theme-and-components.md,references/18-mcp-apps/view-react/09-useviewtool.md - ChatGPT apps:
references/18-mcp-apps/chatgpt-apps/01-dual-protocol.md,references/18-mcp-apps/chatgpt-apps/02-legacy-window-openai-and-skybridge.md,references/18-mcp-apps/chatgpt-apps/03-csp-differences.md,references/18-mcp-apps/chatgpt-apps/04-runtime-detection.md - Next.js:
references/19-nextjs-drop-in/01-overview-withmcpuse.md,references/19-nextjs-drop-in/02-route-and-file-placement.md,references/19-nextjs-drop-in/03-views-in-nextjs.md,references/19-nextjs-drop-in/04-deploying-on-vercel.md - Inspector:
references/20-inspector/01-overview.md,references/20-inspector/02-cli.md,references/20-inspector/03-connection-settings.md,references/20-inspector/04-url-parameters.md,references/20-inspector/05-keyboard-shortcuts-and-palette.md,references/20-inspector/06-integration-and-add-to-client.md,references/20-inspector/07-self-hosting.md,references/20-inspector/08-debugging-chatgpt-apps.md,references/20-inspector/09-changelog-pointer.md - Tunneling:
references/21-tunneling/01-overview.md,references/21-tunneling/02-when-to-tunnel-and-debugging.md - Validate:
references/22-validate/01-inspector-walkthrough.md,references/22-validate/02-curl-handshake.md,references/22-validate/03-connect-real-clients.md,references/22-validate/04-unit-testing-server-fetch.md - Debug:
references/23-debug/01-debugging-workflow.md,references/23-debug/02-transport-debugging.md,references/23-debug/03-view-debugging.md - Production:
references/24-production/01-env-config.md,references/24-production/02-error-strategy.md,references/24-production/03-health-and-custom-routes.md,references/24-production/04-security-hardening.md,references/24-production/05-scaling-stateless.md - Deploy:
references/25-deploy/01-decision-matrix.md,references/25-deploy/02-pre-deploy-checklist.md,references/25-deploy/03-docker.md,references/25-deploy/04-cli-and-org-management.md,references/25-deploy/platforms/01-mcp-use-cloud.md,references/25-deploy/platforms/02-vercel.md,references/25-deploy/platforms/03-cloudflare-workers.md,references/25-deploy/platforms/04-google-cloud-run.md,references/25-deploy/platforms/05-supabase.md,references/25-deploy/platforms/06-deno.md,references/25-deploy/platforms/07-bun.md,references/25-deploy/platforms/08-hono.md,references/25-deploy/platforms/09-railway.md,references/25-deploy/platforms/10-runtime-patterns.md - Anti-patterns:
references/26-anti-patterns/01-sdk-misuse.md,references/26-anti-patterns/02-tool-design.md,references/26-anti-patterns/03-schemas.md,references/26-anti-patterns/04-results.md,references/26-anti-patterns/05-security-and-cors.md - Troubleshooting:
references/27-troubleshooting/01-error-catalog.md,references/27-troubleshooting/02-quick-diagnostic-table.md,references/27-troubleshooting/03-oauth-issues.md,references/27-troubleshooting/04-view-rendering-issues.md,references/27-troubleshooting/05-csp-violations.md,references/27-troubleshooting/06-decision-tree.md - Migration:
references/28-migration/01-from-modelcontextprotocol-sdk.md,references/28-migration/02-v1-to-v2-overview.md,references/28-migration/03-v1-to-v2-imports-server-and-tools.md,references/28-migration/04-v1-to-v2-responses-and-helpers.md,references/28-migration/05-v1-to-v2-auth.md,references/28-migration/06-v1-to-v2-widgets-to-views.md,references/28-migration/07-v1-to-v2-sessions-transports-stdio-sse.md,references/28-migration/08-appssdk-to-mcp-apps.md - Templates:
references/29-templates/01-overview-and-decision-matrix.md,references/29-templates/02-template-mcp-server.md,references/29-templates/03-template-mcp-apps.md,references/29-templates/04-template-blank-and-manual.md - Workflows:
references/30-workflows/01-greenfield-tool-server-to-vercel.md,references/30-workflows/02-views-app-chart-widget.md,references/30-workflows/03-oauth-protected-server-clerk.md,references/30-workflows/04-supabase-oauth-and-deploy.md,references/30-workflows/05-nextjs-drop-in.md,references/30-workflows/06-openapi-to-mcp.md,references/30-workflows/07-proxy-gateway.md,references/30-workflows/08-elicitation-input-required-flow.md - Canonical examples:
references/31-canonical-examples/00-how-to-use-this-cluster.md,references/31-canonical-examples/01-chart-builder.md,references/31-canonical-examples/02-diagram-builder.md,references/31-canonical-examples/03-example-inventory.md
Installationen
Sicherheitsprüfung
Quellcode ansehen
yigitkonur/skills-by-yigitkonur
Mehr aus dieser Quelle
Give agents clean web context
Search and extract the public web as Markdown or structured JSON through one API or hosted MCP server.
Verwandte Skills
Mehr aus dieser Quelle: yigitkonur/skills-by-yigitkonur
So verwenden Sie diesen Skill
Install build-mcp-use-server by running npx skills add yigitkonur/skills-by-yigitkonur --skill build-mcp-use-server in your project directory. Führen Sie den obigen Installationsbefehl in Ihrem Projektverzeichnis aus. Die Skill-Datei wird von GitHub heruntergeladen und in Ihrem Projekt platziert.
Keine Konfiguration erforderlich. Ihr KI-Agent (Claude Code, Cursor, Windsurf usw.) erkennt installierte Skills automatisch und nutzt sie als Kontext bei der Code-Generierung.
Der Skill verbessert das Verständnis Ihres Agenten für build-mcp-use-server, und hilft ihm, etablierte Muster zu befolgen, häufige Fehler zu vermeiden und produktionsreifen Code zu erzeugen.
Was Sie erhalten
Skills sind Klartext-Anweisungsdateien — kein ausführbarer Code. Sie kodieren Expertenwissen über Frameworks, Sprachen oder Tools, das Ihr KI-Agent liest, um seine Ausgabe zu verbessern. Das bedeutet null Laufzeit-Overhead, keine Abhängigkeitskonflikte und volle Transparenz: Sie können jede Anweisung vor der Installation lesen und prüfen.
Kompatibilität
Dieser Skill funktioniert mit jedem KI-Coding-Agenten, der das skills.sh-Format unterstützt, einschließlich Claude Code (Anthropic), Cursor, Windsurf, Cline, Aider und anderen Tools, die projektbezogene Kontextdateien lesen. Skills sind auf Transportebene framework-agnostisch — der Inhalt bestimmt, für welche Sprache oder welches Framework er gilt.