Course Outline
From autocomplete to agents: understanding agent failures
• Anatomy of a coding agent: model, harness, tool surface, context, permissions
• Positioning of tools: Claude Code, GitHub Copilot, Cursor, Codex CLI, Gemini CLI
• A taxonomy of failure modes: incorrect context, mismatched tools, lack of feedback, unbounded autonomy
Demonstration: The same task executed successfully and poorly, side by side
Context engineering
• The context window as a budget: what earns a place within it
, CLAUDE.md, .cursor/rules, copilot-instructions.md — one concept, multiple filenames, single source of truth
• Conventions, build and test commands, architectural boundaries
• <--[endif]-->Retrieval versus explicit context; task decomposition and sub-agents
Lab: Write repository context for an unfamiliar Python service, then re-run a failing task and compare the output
Reusable workflows and Agent Skills
• <--[endif]-->Choosing the abstraction: instruction file, skill, custom command, or plain script
• <--[endif]-->Anatomy of a skill: triggering, instructions, bundled scripts, progressive disclosure
• <--[endif]-->Portability across tools and where vendor lock-in begins
• <--[endif]-->Versioning, review, and distribution across a team; common anti-patterns
Lab: Build and test a reusable workflow that enforces a house coding standard
MCP: connecting agents to real systems
• <--[endif]-->Architecture: clients, servers, tools, resources, prompts; stdio and HTTP transports
• <--[endif]-->Servers that earn their place: Git hosting, issue trackers, databases, browsers, internal APIs
• <--[endif]-->When a CLI or script outperforms an MCP server
• <--[endif]-->Tool-surface hygiene: why more tools lead to less reliability
Lab: Wire up MCP servers and handle a ticket end-to-end — issue, branch, patch, tests, pull request
Feedback loops and evaluation
• <--[endif]-->Tests, types, and linters as the agent’s ground truth; test-first work as a control mechanism
• <--[endif]-->CI as the outer loop, and review discipline for agent-authored diffs
• <--[endif]-->Golden-task evaluation sets: what to measure and how to catch regressions
• <--[endif]-->Cost and latency as first-class metrics
Lab: Build a small evaluation set and score two agent configurations against it
Security and guardrails
• <--[endif]-->Prompt injection through issues, pull requests, READMEs, dependencies, and fetched pages
• <--[endif]-->Permission models: allowlists, approvals, read-only tools, network egress control
• <--[endif]-->Secret hygiene and sandboxing: containers, ephemeral credentials, limiting blast radius
• <--[endif]-->Supply-chain risk in third-party MCP servers and shared skills
Lab: Watch an agent get hijacked by a poisoned repository, then harden the setup to prevent it
Rolling this out to a team
• <--[endif]-->A staged adoption path; what to standardize and what to leave to individuals
• <--[endif]-->Metrics that indicate real value, and those that do not
Requirements
• Working knowledge of Python, Git, and the command line
• Prior exposure to an AI coding assistant
• <--[endif]-->NobleProg will provision Dadesktop VMs for participants with Docker, VS Code, and Python 3.11 or later
• A functional AI coding assistant of the participant’s choice: Claude Code, GitHub Copilot, Cursor, Codex CLI, or Gemini CLI. The labs are tool-agnostic, with instructions provided for each option
Audience
• Software engineers, technical leads, and architects who use AI coding assistants but struggle to achieve reliable results
• Platform and developer-experience engineers deploying AI tooling across teams
• Engineering managers establishing standards, guardrails, and success metrics
Testimonials (3)
The trainer is patient and very helpful. He knows the topic well.
CLIFFORD TABARES - Universal Leaf Philippines, Inc.
Course - Agentic AI for Business Automation: Use Cases & Integration
Good mixvof knowledge and practice
Ion Mironescu - Facultatea S.A.I.A.P.M.
Course - Agentic AI for Enterprise Applications
The mix of theory and practice and of high level and low level perspectives
