Azzist
End-to-end project shipping: scaffold a stack, dockerize, deploy over SSH into an isolated Docker network, wire a reverse proxy (nginx or existing Traefik), and attach a public domain via Cloudflare — all without disturbing existing server services. Use when the user wants to build AND deploy something live, says "azzist", "ship this e2e", "deploy a landing page", "make it public", or asks for a full develop->deploy flow.From its SKILL.md
npx -y skills add ghulammuzz/azzist-skills --skill azzistAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
2 things to look at
- no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
- 1 stars1 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
3.3 KB, as published. Nobody here has run it
azzist (orchestrator)
Drive a project from idea to public URL by running the focused sub-skills in order. Each stage has its own skill; invoke them as you reach that stage.
Stack conventions (do not deviate unless the user asks)
| Need | Stack |
|---|---|
| Web frontend / landing | Nuxt 3 |
| Backend API | Go |
| Worker / background jobs | Go + GoFiber |
| Data-related work | Python (tidy structure) |
| Database | Postgres |
| Cache / queue | Redis |
| Packaging | Dockerfile per service |
| Reverse proxy | nginx (or existing Traefik) |
v1 implements the Nuxt web golden path end-to-end. Other stacks are scaffold-planned.
Flow
First decide: new project or existing project?
- Existing code present (source files already in the dir/repo) → run azzist-analyze
to comprehensively trace the codebase, detect the stack/port/services, and propose an
azzist.yaml. Then go to step 1 (Config) and skip Scaffold. - Empty / new → follow Config → Scaffold → Deploy → Server.
- Config — if
azzist.yamlis absent, run azzist-init to createazzist.yaml(committed) andazzist.local.yaml(gitignored secrets). For an existing project, reuse theazzist.yamlproposed by azzist-analyze and only generate the secrets file. Make sure the user has filled the domain, Cloudflare token/zone, and server SSH details before any remote step. - Scaffold (new projects only) — run azzist-scaffold to generate the project and
confirm
npm run devworks locally. - Deploy — run azzist-deploy to build the image, smoke-test it, and run it isolated
on the server (
azzist_<name>network, private port). Pause and confirm with the user before this first remote/destructive step (it installs Docker if missing and runs containers on their server). - Server — run azzist-server to wire the proxy non-invasively (Traefik labels or an isolated nginx vhost), upsert the Cloudflare A record, and provision TLS.
Isolation contract (applies to every stage)
- Touch only resources named
azzist_<name>*. Never stop, remove, or reconfigure existing containers, networks, proxies, or services. If something unexpected blocks you, ask the user instead of forcing past it. - Existing Traefik is integrated via labels only — never reconfigured.
- Secrets stay in
azzist.local.yaml; never commit them.
E2E success check
The flow is done when:
curl -fsS -I https://<domain>
returns 200 (or 301 to https) and the site is publicly reachable. Report the public URL,
the proxy strategy used, the chosen port, and the Cloudflare record result.