Skip to main content
skill CX Codex

wendy-cloud-iterate — Claude Code skill: Wendy Cloud test loop

Runs the Wendy Cloud dev stack end to end: brings up Docker and the Swift broker, runs integration tests, reads the failures and edits code.

Install into Claude Code · Skills CLI (Vercel) All methods and platforms ↓
npx skills add wendylabsinc/claude-skills --skill wendy-cloud-iterate -a claude-code
Open on GitHub

What wendy-cloud-iterate does

Around the Wendy Cloud repository this skill closes the loop by itself. It inspects the machine first, starts the stack, executes the test suite, works out what broke and applies a fix. Then it decides when to go round again. Two entry points exist: make dev-docker keeps the broker inside Docker, while make dev-local builds it on the host for quicker rebuilds. Both bring up Postgres, swift-broker, tunnel-broker and the gRPC-web Envoy, and seed a dev user with an org.

The prerequisite block is the interesting part. It installs Homebrew, gh, protoc and the wendy CLI, starts Docker Desktop if the daemon is down, clones wendylabsinc/pki-core and wendylabsinc/cloud into ~/Documents/Projects, and checks GitHub CLI auth. After that no confirmation prompts appear: the skill pulls repos, runs make, docker compose and wendy commands, edits sources, commits and raises a pull request with gh pr create.

Two things can stop it. wendy auth login relies on Chrome and the Claude browser extension; without them you finish the login yourself — open the printed URL, pick an org, wait until the certificates are saved. And that login does not currently authenticate against the broker: a session appears, yet data-bearing RPCs answer UNAUTHENTICATED. The repo tracks this as WDY-2330, so the auth steps show the shape of the flow rather than a route to a working certificate.

What it helps with

  • bring up the Wendy Cloud stack in one command
  • run integration tests and triage the failures
  • track down a regression after a new feature
  • verify the toolchain before building the broker
  • start Docker and wait until the daemon answers

Example prompts for the agent

  • Run the Wendy Cloud test loop and fix whatever fails
  • Start the stack with make dev-docker and confirm the broker ports are listening
  • Scan the Swift broker for regressions after the latest changes

How to install wendy-cloud-iterate

1Skills CLI (Vercel)

npx skills add wendylabsinc/claude-skills --skill wendy-cloud-iterate -a claude-code

2Manually: put the files in place

~/.claude/skills/wendy-cloud-iterate/SKILL.md

Copy the whole skill folder; for one project use .claude/skills/wendy-cloud-iterate/

Step by step, with commands for each platform: How to install skills in Claude Code: five ways

Questions about wendy-cloud-iterate

How do I install wendy-cloud-iterate in Claude Code?

Run npx skills add wendylabsinc/claude-skills --skill wendy-cloud-iterate -a claude-code. The skill lands in ~/.claude/skills/wendy-cloud-iterate/SKILL.md.

What does wendy-cloud-iterate do?

It drives an autonomous CI loop on the Wendy Cloud repo: starts the dev stack, runs tests, diagnoses failures, edits code and opens a pull request.

Why can the wendy CLI login fail?

CLI auth against the local stack is broken end to end (WDY-2330): the dev cookie yields a session, but RPCs that carry data return UNAUTHENTICATED.

Similar skills

All skills for Testing and QA →

From the same set7

Project sponsors

Companies that keep this analytics open
Gold sponsor

Impressions, clicks and code copies — open statistics. We keep the numbers public for the same reason NeBlask itself is public. Become a sponsor