jobber-cli

jobber-cli

Agent-first private-surface CLI for Jobber.

This package is intentionally v0:

• browser-assisted auth bootstrap
• raw GraphQL execution
• current user / account identity checks
• generated operation inventory from authenticated frontend bundles
• read-first defaults

Agent Quickstart

From a live logged-in Jobber browser session on the same machine:

jobber auth import-agent-browser
jobber doctor --json
jobber whoami --json
jobber operations list --search client --json

For raw one-off GraphQL without creating temp files:

jobber graphql run \
  --operation-name CurrentAccount \
  --query 'query CurrentAccount { account { id name inTrial industry } }' \
  --json

Install

Local:

git clone https://github.com/danielgwilson/jobber-cli.git
cd jobber-cli
npm install
npm link

Global:

npm install -g jobber-cli

Auth

Preferred from a live logged-in agent-browser session:

jobber auth import-agent-browser
jobber auth status --json

Manual auth:

printf '%s' "$JOBBER_COOKIE_HEADER" | jobber auth set-cookie-header --stdin
jobber auth status --json

Saved config lives at ~/.config/jobber/config.json with 0600 permissions.

Supported env vars:

• JOBBER_COOKIE_HEADER
• JOBBER_GRAPHQL_VERSION
• JOBBER_API_URL

Skill Install

Install the canonical nested skill:

npx -y skills add -g danielgwilson/jobber-cli --skill jobber

The skill lives at skills/jobber/SKILL.md. The npm package includes that nested skill and its Codex metadata; it does not publish a duplicate root SKILL.md.

Main Commands

jobber doctor --json
jobber whoami --json
jobber auth status --json
jobber operations list --json
jobber operations list --search client --json
jobber operations inspect ClientDefaults --json
jobber graphql run --operation-name CurrentAccount --query 'query CurrentAccount { account { id name inTrial industry } }' --json
jobber graphql run --query-file ./query.graphql --variables-file ./vars.json --json

Notes

• This adapter depends on Jobber's private web surface and is therefore fragile.
• operations list is a discovered inventory, not a guarantee that every operation has a recovered query document yet.
• graphql run is intentionally generic so new workflows can be tested before typed subcommands are added.
• auth import-agent-browser assumes a live authenticated agent-browser session or attached browser on the same machine.