vercel-cli-with-tokens

Vercel CLI with Tokens

Deploy and manage projects on Vercel using the CLI with token-based authentication, without relying on vercel login. The token is resolved once, held outside the conversation and the repo, and the project/team binding is remembered between runs.

This skill bundles its own steer runtime at scripts/steer.py; the commands below invoke it with python3 and need nothing installed (the Vercel CLI itself excepted). Paths are relative to this skill's directory: when your working directory is elsewhere (it usually is), use the skill's full path (python3 <path-to-this-skill>/scripts/steer.py ...).

Before you start

1. Ground yourself. Run python3 scripts/steer.py context --tools vercel,node,npm and read the snapshot; it tells you the platform, git state (including whether a remote exists), and whether the Vercel CLI is installed. If the CLI is missing: npm install -g vercel.
2. Apply past lessons. Run python3 scripts/steer.py learn show and follow what it says; those lessons came from real previous runs.

Step 1: Resolve the token

Never read a token value into the conversation; check by name, export by substitution. Work through these in order:

1. Ask steer. python3 scripts/steer.py secrets check VERCEL_TOKEN resolves the environment, the OS keychain, and steer's own secret store in one command. If present, make it available to the CLI:

   export VERCEL_TOKEN="$(python3 scripts/steer.py secrets get VERCEL_TOKEN)"
   

2. A .env file may hold it. Detect variable NAMES only; do not print values (Vercel tokens look like vca_...):

   grep -o '^[A-Za-z_]*VERCEL[A-Za-z_]*=' .env 2>/dev/null
   export VERCEL_TOKEN="$(grep '^<VARIABLE_NAME>=' .env | cut -d= -f2-)"
   

   Then offer to persist it for future runs so this discovery never repeats: ask the user to run python3 scripts/steer.py secrets set VERCEL_TOKEN (hidden prompt, lands in the keychain or a 0600 file, never in the repo or the chat).