Claude Code
Anthropic's terminal-based coding agent. Reads and writes across your repo, runs your tests, uses MCP tools, executes shell commands under your permission. The thing that actually writes code.
The engagement runs as three stages that match the offer. Stages 1 and 2 are the step. Stage 3 is the stride. Stride itself, the software, stays complimentary for as long as you use it.
Two days that install the tools, teach the primitives, and ship the first app end to end.
The workflow only becomes real when you're building alone. I book one day each week to catch anti-patterns while the habits are still forming.
Past three weeks, support drops to what you actually need. Stride itself stays complimentary.
A working app on day one. A week of toolchain decisions already made. Boring, opinionated, proven.
Anthropic's terminal-based coding agent. Reads and writes across your repo, runs your tests, uses MCP tools, executes shell commands under your permission. The thing that actually writes code.
A thin wrapper around Claude Code that implements kanban on Linear. Plain-English brief in, well-scoped Linear card out. Branches cut, atomic commits landed, PRs opened. The bit you don't have to think about.
Fast, modern issue tracker. Jira's sleeker cousin. Your work is visible without you having to write a status report. One URL to your line manager and they see everything.
A working web app or Python scaffold from the first commit. So you don't spend three days picking a stack. React + Vite + ShadCN + Biome + tests for web. Tooling pre-wired for Python.
One human orchestrating AI agents is not a classic team. Agents have no commitment cost and they work in parallel, so sprint boundaries become artificial lag. The real WIP caps sit at review bandwidth and merge coherence. Both are bottlenecked on you, the human.
Kanban handles that natively. Cards pull through lanes when there is capacity. No sprints, no ceremonies, no planning meeting, no retro that nobody wants.
No. On a Claude Team seat, prompt-training prohibition is contractual. On a personal Max subscription, it's a toggle you own. Either way, the answer is no.
Your GitHub org. Your accounts. Your infrastructure. Webventurer does not host anything. If you leave Webventurer, nothing is locked in. Everything keeps working in place.
The full procurement brief covers DPA, the private GitHub org, the Claude Max subscription or 5+ Team seats requirement, and Anthropic's Commercial Terms. Shareable with InfoSec and your DPO.
Around £6k ex VAT. £5k is my engagement. The rest is Anthropic and Linear subscriptions and hosting. Three replaced vendor reports pay it back. Every report after is upside.
30 minutes, free. We walk through your actual data sources and map them to Linear cards before any money is spent. You leave with a one-page scope.