Programme that teaches main concepts on agentic development and acts as a coach in your claude code sessions.
The picture of you, not of the topic. Each statement is something you'll be able to do by the end.
I can Implement and Maintain Claude Development Best Practices
I can Design and Coordinate Multi-Agent Systems
I can Evaluate and Optimize Claude Code Performance
I can Apply Reasoning to Claude Development Decisions
The path through the material. Each lesson tackles one essential question.
What foundations turn a fresh install into an agent that speaks our language, respects team norms, and delivers against a clear goal?
How do we equip the agent and shape its execution so it can do real work safely, economically, and within a clearly bounded scope?
When tasks grow, how do planning, parallel work, and key integrations keep progress flowing without micromanagement?
How can we bake conventions, diagnostics, and structure into the environment so quality emerges by default instead of after-the-fact fixes?
What practices keep sessions coherent, recoverable, and extensible as we juggle teams, repositories, and evolving toolchains?
How do we extend the agent responsibly and exercise engineering judgment so errors become learnings rather than liabilities?
When an agent derails, what disciplined moves restore a clean state so we can continue safely instead of compounding errors?
How do we run and embed agents as predictable, bounded components we can automate and integrate?
How do we frame failures and evidence so the agent accelerates tight test-fail-fix loops and isolates causes without exploring the whole codebase?
Why do concrete constraints and the right reasoning budget let the agent produce trustworthy reviews and edits instead of vague approximations?
How can we enforce formatting, secret hygiene, and path protections automatically so the agent operates safely in local runs and CI?
How do executable intent—specs and tests—turn the agent into a reliable implementer and keep large refactors reversible?
Three ways to connect: Claude Code (PAT + install command), Claude Desktop (.mcpb download — no token to paste), or Claude web (Customise → Connectors → Add custom connector, OAuth). Same MCP endpoint, same identity on every path.
https://nebular.live/api/v1/mcp/