Claude Code setup & support - Claude Code, configured for the way your team actually works.
Most teams buy Claude Code seats and use 10% of what it can do. We install the slash commands, hooks, MCP servers, subagents, and permissions your engineers need — packaged as a private plugin so everyone onboards in 60 seconds. Two weeks end to end.
What we install - Six building blocks. Built once, used by the whole team.
Every engagement ships these six pieces. Mix and match — most teams use four or five of the six. All wrapped in a private plugin so updates go out to every engineer at once.
- Custom slash commands. Codified workflows — /ship, /review, /scaffold-test, /security-pass — wired to your repo conventions so engineers stop re-typing the same prompt 30 times a week.
- Hooks. PreToolUse, PostToolUse, Stop, and SessionStart hooks that enforce your standards: lint before commit, route bash commands through your sandbox, block writes outside the repo, ping Slack on long jobs.
- MCP servers. Custom Model Context Protocol servers that expose your Linear, Notion, internal APIs, monitoring, and databases to Claude Code with proper auth and scoping.
- Subagents. Specialist subagents — explore, code-reviewer, security-auditor, build-validator — defined to your engineering standards so Claude delegates the right work to the right pattern.
- Settings & permissions. Per-role settings.json templates: tight permission allowlists for new joiners, broader access for seniors, audited bash commands, and shared environment configuration.
- Plugins for your org. A private plugin bundling everything above, distributed across your team via your own marketplace — every engineer onboards with the same setup in 60 seconds.
Why bother - What changes for your engineers.
The point isn't fancier prompts. It's an environment where Claude Code defaults to your patterns, your standards, and your tools — without anyone having to remember to set it up.
- Faster PR cycle. Pre-baked review and test scaffolding means PRs land cleaner the first time. Less back-and-forth, less context-switching for senior engineers.
- Onboarding from days to hours. New hires get a configured Claude Code from day one with the same hooks, commands, and access patterns as the rest of the team.
- Standards enforced, not documented. Your "we always do X" engineering rules become hooks and subagents — enforced by tooling, not by hoping engineers read the wiki.
- Secrets and prod stay safe. Permission allowlists, sandboxed bash, and audit logs mean Claude Code can move fast on the codebase without putting credentials or production in reach.
How we work - Two weeks. Founder-led. Your repo, your cloud.
Same shape as our other engagements — short, fixed scope, real output inside the first week. We work inside your repo and your accounts; nothing of yours leaves your environment.
- Day 1 — Discovery. One call with your engineering lead to map the workflows worth codifying, the systems Claude Code needs MCP access to, and the permission shape your team needs.
- Day 7 — First prototype. Three slash commands and one MCP server installed in your repo, tested by you. You see the value before paying the second invoice.
- Day 14 — Plugin shipped. Full plugin published to your private marketplace, installation doc handed over, every engineer onboarded. We stay on for 30 days of post-launch support included.
FAQ - The questions engineering leads ask first.
If yours isn't here, the contact form is the fastest way to ask.
- Does this work with Cursor / Copilot / Cody?. We focus on Claude Code because that's where most of the agentic capability lives right now (hooks, subagents, MCPs, the plugin system). Some pieces — MCP servers in particular — work with any MCP-capable client, so the work isn't wasted if you change tools later.
- Do you need access to our repo?. For the build itself, yes — we work in a feature branch you grant us access to. For ongoing support, no; we can review configs and ship updates against a fork if your security policy prefers that.
- Where do credentials live?. In your environment. MCP servers run in your infrastructure with your secrets. We never store API keys, OAuth tokens, or credentials of any kind on our side.
- Is the plugin distributed publicly?. No — by default it lives in a private marketplace under your org. Only your engineers see and install it. We can also ship to the public Claude Code plugin marketplace if you have a reason to.
- Do you train our team or just install?. Both. The handover is two-part: written docs covering every command, hook, and MCP we install — plus a 60-minute live session with your engineers walking through the workflows.
- What about ongoing changes?. Optional monthly retainer. Claude Code itself ships fast, your stack changes, new MCPs become useful. Most teams want a few hours a month to keep the plugin current — that's the Growth Retainer on the pricing page.
Let's build something great together.
Melbourne, Australia — serving founders worldwide. [email protected]