Claude Code workflows that survive a real week
Claude Code is a terminal coding agent from Anthropic. It works best when the team designs the workflow around it.
What this covers
What I set up and check with your team.
Workflow first, prompt second
Feature work, refactors, bug triage and docs are mapped first. Prompt patterns come after the workflow is clear.
Review checks and slash commands
Cheap checks, formatters and validators catch common mistakes. Slash commands cover the few recurring prompts worth standardising.
Background research with bounds
Research, audits, dependency scans and test triage stay bounded so the main session stays on the change.
MCP servers where they earn their keep
Private docs, internal APIs and query bridges are wired only where the agent genuinely needs them.
Memory and context discipline
CLAUDE.md and feature context stay deliberate: what belongs in memory, what belongs in the prompt, and what belongs nowhere.
Where the checks sit
Developers already use it
Nearly half of developers use AI tools every day. The question is what the tool is allowed to change.
- Daily47.1%
- Weekly17.7%
- Monthly or less13.7%
- Not yet, planning to5.3%
- No plans16.2%
Where AI code goes wrong
The most common complaint is output that is almost right. Hooks and tests catch that before it lands.
- Almost right, but not quite66%
- Debugging AI code takes longer45.2%
- Less confident in own problem-solving20%
- Hard to understand how the code works16.3%
See the engagement shape
Review the sequence, review gates and handover before you book.
How the work runs
A bounded sequence turns the tool decision into an operating habit.
- Day 0
Read-only review
Repo, CI, and current developer workflow are mapped. We agree the first checks, commands, and MCP servers worth wiring up.
- Day 1
Workflow design
Workflow design, checks, slash commands, MCP short list and review patterns.
- Day 2
On the real repo
Real Claude Code sessions on agreed tasks. Checks run, diffs land in review.
- Weeks 2 to 4
Rollout window
Light coaching, an evaluation suite carried by the team, written rules of engagement, and a check-in to confirm the workflow holds in real weeks.
Questions teams ask before the work
How is this different from training people on ChatGPT?
Claude Code lives in the terminal and reads the repository. The workshop is about workflow design and review discipline around an agent that has access to the codebase, not about prompt phrasing for a chat box.
Do you build custom MCP servers?
Yes, where the team has a genuine internal tool worth surfacing. MCP server design is a separate service, described on the MCP page of this site. On the workshop we focus on wiring up servers that already exist or that the team can build in a day.
Will this work in a regulated environment?
It is built for one. Read-only access first, named review gates, policy checks, and the M365 and identity layer underneath all stay part of the engagement, not an afterthought.
Keep exploring
Book a Claude Code workshop
Best for teams that already use a CLI seriously — git, Make, language toolchains. Less suited to teams who live entirely in a UI editor and want a one-click integration.