Advanced Context Engineering
Focusadvanced prompting techniques and when each applies; RePPIT (Research, Propose, Plan, Implement, Test) and spec-driven development; MCP fundamentals (servers, clients, tools, transport); designing tools for agent ergonomics.
You buildA one-page spec, one feature shipped through RePPIT, and an MCP server with 2–4 tools
Core material≈ 3 h 50 min
- 1Video/podcast1 h 33 minContext EngineeringThe Pragmatic Engineer with Dex Horthy
Context, harnesses, loops, research/plan/implement workflows, compaction, and software factories. The article and transcript are useful for notes.
- 2Article25 minEffective context engineering for AI agentsAnthropic
The canonical framing of context as a finite resource: system prompts, tools, examples, retrieval, and compaction.
- 3Video29 minPrompting for AgentsAnthropic
How prompting changes when the model runs in a loop with tools.
- 4Article25 minWriting effective tools for agentsAnthropic
Tool names, descriptions, boundaries, results, and evaluation.
- 5Specification30 min for the architecture, transports, and tools sectionsSpecificationModel Context Protocol
Read the primary source before any tutorial.
- 6DocumentationSpec KitGitHub
And its Agentic SDD workflow (20 min) — a concrete implementation of spec-driven development.
Deeper material
- Courseunits 0–2, roughly 6–8 h; spread it across Weeks 2 and 3MCP CourseHugging Face
Free, hands-on foundations, continuing into the end-to-end MCP application unit.
- LinkContext Engineering for AI Agents: Lessons from Building ManusManus
KV-cache hit rate, file system as context, and keeping failures in the trace.
- LinkHow Long Contexts FailDrew Breunig
And Chroma, Context Rot — the evidence that more context is not free.
- LinkSpecs Are the New Source CodeRavi Mehta
And the Kiro specs docs — the product side of spec-driven development.
- Video10 minHow I use Claude Code for real engineeringMatt Pocock
Planning, clarifying questions, phased execution, and context-window management.
Additional video track
- Advanced prompting1 h 16 minAI Prompt Engineering: A Deep DiveAnthropic
, and the shorter Prompting 101 (24 min).
- What still works1 h 37 minAI prompt engineering in 2025: What works and what doesn'tSander Schulhoff on Lenny's Podcast
Evidence-based prompting, few-shot, decomposition, and prompt injection.
- Research, plan, implement21 minNo Vibes Allowed: Solving Hard Problems in Complex CodebasesDex Horthy
Targeted research, deliberate compaction, concrete plans, and human review.
- Long-horizon context39 minContext Engineering Our Way to Long-Horizon AgentsHarrison Chase of LangChain
.
- Spec-driven development1 h 04 minSpec-Driven Development: Agentic Coding at FAANG Scale and QualityAl Harris of Amazon Kiro
Requirements, acceptance criteria, design artifacts, tasks, and steering documents.
- MCP workshop1 h 44 minBuilding Agents with Model Context ProtocolMahesh Murag of Anthropic
The full workshop from the protocol's authors.
- MCP video course≈ 1 h 30 minMCP: Build Rich-Context AI AppsDeepLearning.AI and Anthropic
Architecture, servers, clients, tools, resources, prompts, remote deployment, and Inspector-based testing.
Build
Choose one real feature for your repository. Write a one-page spec containing objective, constraints, out-of-scope items, acceptance checks, edge cases, and a test plan. Follow Research → Propose → Plan → Implement → Test without skipping a stage, and keep the artifact from each stage.
Then build one MCP server with two to four narrow tools. Exercise it in MCP Inspector first, then from your agent. Test happy paths, invalid parameters, huge results, timeouts, and permission failures. Revise each description until a fresh agent consistently selects the right tool and passes valid arguments.