Developing in Agentic AI Systems (GH-600)

Foundations of Agentic AI in GitHub

In this section, learners explore how AI coding agents are changing software development by planning, acting, and improving within GitHub workflows.

• Define agentic AI in the SDLC
• Explain the agent lifecycle: plan, act, evaluate
• Describe GitHub as the system of record and control plane
• Identify responsibilities, risks, anti-patterns, and traceability needs
• Apply the contributor model to agent-generated work

Agent Architecture and SDLC Integration

In this section, learners design agentic systems that use GitHub workflows to build software safely, reliably, and with appropriate controls.

• Map agent responsibilities to the SDLC
• Define inputs, outputs, and success criteria
• Separate planning, reasoning, and execution
• Apply pull request governance using templates, checks, CODEOWNERS, rules, and environment gates
• Build reliable workflows with outputs, contexts, triggers, and cross-job handoffs
• Control and operate agents using observability, tools, MCP, secrets, hooks, and reliability practices

Tooling, MCP, and Agent Execution Environments

In this section, learners configure the tools, execution environments, and boundaries agents need to perform tasks safely within GitHub workflows.

• Explain how agents interact with GitHub APIs and workflows
• Use Model Context Protocol servers, registries, and allow lists
• Define execution context and boundaries
• Apply agent execution limits and protections

Multi-Agent Systems and Orchestration

In this section, learners design reliable multi-agent systems in GitHub using observable workflows, coordinated artifacts, and safe recovery mechanisms.

• Design multi-agent workflows for coordinated development tasks
• Define agent roles, responsibilities, and handoff points
• Coordinate artifacts across agents and workflows
• Monitor multi-agent activity for reliability and traceability
• Apply recovery mechanisms when agent workflows fail or require intervention

Memory, State, and Evaluation

In this section, learners manage agent memory and state, persist progress across environments, and evaluate agent behaviour using clear success signals.

• Manage agent memory and state across tasks
• Persist progress across development environments
• Define evaluation criteria for agent outputs
• Use scans, artifacts, and signals to assess quality
• Tune agent behaviour based on evaluation results and error analysis

Governance, Guardrails, and Operations

In this section, learners design secure and compliant agent governance using GitHub-native controls, human approvals, least-privilege access, and operational safeguards.

• Implement governance and guardrails for agentic workflows
• Configure human-in-the-loop approvals
• Apply least-privilege access controls
• Improve accountability through logging, review, and traceability
• Use operational safeguards to support reliability, recovery, and safe production use