Software notes for teams building real systems.
Browse by workstream, open a focused guide, or search the library when you already know what you need.
Choose the workstream that matches the problem.
Browse practical notes by topic and need.
- AI Agent Context Budget: Allocate Tokens by Evidence ValueArtificial Intelligence
- Agent Context Provenance: Show Where Every Instruction and Fact Came FromArtificial Intelligence
- Structured Outputs for AI Agents: Version, Validate and RecoverArtificial Intelligence
- LLM Gateway Build vs Buy: A Control-Plane Decision for Enterprise AIArtificial Intelligence
- AI Model Routing: Match Each Agent Step to the Right ModelArtificial Intelligence
- AI Agent Evaluation Scenarios: Test the Workflow, Not Just the AnswerArtificial Intelligence
- OpenTelemetry AI Agent Tracing: Models, Tools and HandoffsArtificial Intelligence
- AI Agent Approval UX: What Reviewers Must See for High-Risk ActionsArtificial Intelligence
- AI Agent Sandbox: Isolation for Code, Files and BrowsersArtificial Intelligence
- Agent Workflow Compensation for Multi-Agent TransactionsArtificial Intelligence