Developing with and for AI agents
Agentic software development, testing and evals, deployment, operations and cost of AI agents — from practice.
Fundamentals
- FundamentalsAgentic Software Development in Teams: Claude Code, Codex and CursorAgentic coding as a team practice: choosing a tool, shared configuration, permissions, review of agent code and rolling out Claude Code, Codex or Cursor.
- FundamentalsAgentOps: Running AI Agents ReliablyAgentOps explained: ownership, tracing, evals, model retirements, incidents and on-premises operation for AI agents once they are in production.
- FundamentalsCode Review with AI Agents: Keeping Quality at Higher ThroughputHow to set up AI code review that holds up: automated pre-checks, a clear focus for human reviewers and rules that survive high agent throughput.
- FundamentalsCost Control for AI Agents and Coding AgentsAI cost management for agents and coding tools: spend limits, LLM gateways, cost allocation and subscription versus pay-per-token, explained for leaders.
- FundamentalsFrom Spec to Code: Spec-Driven Development with AgentsSpec driven development for coding agents: spec, plan, tasks and verification as one workflow, explained with Spec Kit, Kiro and Claude Code.
- FundamentalsRoles in an AI Agent Team: Who Does WhatAI team roles for mid-sized companies: the four roles you need to build and run AI agents, who can take them on and which ones can be covered externally.
- FundamentalsTesting AI Agents: From Unit Tests to Eval SuitesHow to test AI agents when outputs vary: a test pyramid of unit tests, evals, integration runs and regression suites, with a practical rollout plan.