Topic / AI-assisted Development
AI-assisted Development
Make generated code easier to understand. Give an agent a system map, a clear change, and a way to check its work.
6 articles available
5 min read
Understand an AI Agent's Code with Architecture and Sequence Diagrams
Investigate a fictional release-note agent by connecting its model loop, read tools, and draft storage to code evidence and one concrete scenario.
5 min read
Write a Diagram-Driven Brief for an AI Coding Agent
Specify revision-bound publication for a release-note agent, including immutable approval snapshots, scope, recovery assumptions, and checks.
5 min read
Model Failure Paths Before an AI Agent Implements Them
Work through lost responses and competing draft revisions to decide which recovery actions are allowed and which evidence is still missing.
5 min read
Implement and Verify an AI-Assisted Change Against a Model
Review a publication change from authorization to durable state and remote delivery, choosing checks that can expose each important failure.
5 min read
Keep Architecture Models and Code Aligned with an Explicit Task
Update a release-note agent's model after a background reconciliation worker is added, while preserving the policy and recording what was verified.
4 min read
Review an AI-generated change with a system map
Turn a generated patch into a change you can explain: one promise, one boundary, and a few observable checks.
Learning paths in this topic
Learning path · 3 proposed modules
System design in practice
A proposed route from tracing one request to reasoning about boundaries, failures, and change. Start with a system you can explain.
2 articles available
Learning path · 2 proposed modules
Work with an AI development agent
A proposed sequence for turning a vague request into a bounded change, then checking the result against the system you meant to build.
1 article available
Learning path · 5 proposed modules
AI-assisted development: from model to verified change
Understand a release-note agent, brief a bounded publication change, analyze its failure paths, verify the implementation, and maintain the model through an explicit task.
5 articles available