agentsclimarketplace

Deploy vercel

Skill AayushMS/deploy-skills/skills/deploy-vercel

Deploy frontend applications to Vercel using CLI and REST API. Use when the user wants to deploy a React, Vite, Next.js, or static frontend to Vercel. Triggers on "deploy to Vercel", "Vercel deploy", "host my frontend", "deploy my app to Vercel", or any task requiring Vercel project creation, environment variables, production deployment, or GitHub auto-deploy connection.From its SKILL.md

Install
npx -y skills add AayushMS/deploy-skills --skill deploy-vercel

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

3.8 KB, 989 tokens by cl100k_base, as published. Nobody here has run it

Deploy to Vercel

Install CLI

npm i -g vercel

Authenticate

vercel login

Verify: vercel whoami

CLI Scope Issue

The Vercel CLI often fails in non-interactive (piped/automated) mode with missing_scope errors, even when --scope is provided. Use the REST API instead for project creation, env vars, and git linking. The CLI works reliably for vercel --prod deployments once the project is linked.

Deployment Workflow

1. Create project via API

VERCEL_TOKEN=$(cat ~/.local/share/com.vercel.cli/auth.json | python3 -c "import sys,json; print(json.load(sys.stdin)['token'])")

# Get team/scope ID first
curl -s "https://api.vercel.com/v2/teams" \
  -H "Authorization: Bearer $VERCEL_TOKEN" | python3 -c "
import sys,json
for t in json.load(sys.stdin).get('teams',[]):
    print(f'{t[\"id\"]}: {t[\"name\"]}')"

TEAM_ID="<team_id_from_above>"

curl -s -X POST "https://api.vercel.com/v11/projects?teamId=$TEAM_ID" \
  -H "Authorization: Bearer $VERCEL_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "my-project",
    "framework": "vite",
    "buildCommand": "cd client && npm install && npm run build",
    "outputDirectory": "client/dist",
    "installCommand": "npm install"
  }'

For non-monorepo projects, omit buildCommand/outputDirectory/installCommand and let Vercel auto-detect.

Save the id field from the response as PROJECT_ID.

2. Set environment variables via API

curl -s -X POST "https://api.vercel.com/v10/projects/$PROJECT_ID/env?teamId=$TEAM_ID" \
  -H "Authorization: Bearer $VERCEL_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "key": "VITE_SERVER_URL",
    "value": "https://my-backend.up.railway.app",
    "target": ["production", "preview"],
    "type": "plain"
  }'

3. Create local project link

mkdir -p .vercel
cat > .vercel/project.json << EOF
{"projectId": "$PROJECT_ID", "orgId": "$TEAM_ID"}
EOF

Add .vercel to .gitignore.

4. Deploy to production

vercel --prod --yes

5. Verify

The deploy output shows the production URL (e.g., https://my-project.vercel.app).

Monorepo Considerations

When the frontend imports from a shared directory (e.g., ../shared/types.ts):

  • Set root directory to . (repo root), not client/
  • Override buildCommand to cd client && npm install && npm run build
  • Override outputDirectory to client/dist
  • The installCommand at root level should trigger postinstall hooks for subdeps

Connect GitHub for Auto-Deploy

curl -s -X POST "https://api.vercel.com/v9/projects/$PROJECT_ID/link?teamId=$TEAM_ID" \
  -H "Authorization: Bearer $VERCEL_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "type": "github",
    "repo": "USER/REPO",
    "productionBranch": "main"
  }'

Verify the response contains a "link" object with "type": "github".

Useful API Endpoints

See references/api_reference.md for the full API reference.

ActionMethodEndpoint
Create projectPOST/v11/projects?teamId=
Set env varPOST/v10/projects/{id}/env?teamId=
Link GitHubPOST/v9/projects/{id}/link?teamId=
List env varsGET/v10/projects/{id}/env?teamId=
Update projectPATCH/v9/projects/{id}?teamId=
Delete projectDELETE/v9/projects/{id}?teamId=

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 325,949. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.