In the news
Using coding agent traces to improve your workflows
Pydantic · Duncan McKinnon · Published · 3 min read
In 30 seconds
- What happened
- Pydantic released guidance on using coding agent traces to monitor, debug, and optimize AI-powered development workflows through observability.
- Why it matters
- Teams building or improving coding agents need visibility into model invocations, tool calls, and token usage to identify inefficiencies and close feedback loops.
- Watch out
- Tracing requires instrumentation across providers and harnesses; not all providers supply complete metadata like cache reads or cost estimates.
- agent
- token
The patterns behind this
- Eval-Driven Development (Agent CI)
- Agent Observability & Tracing
- Agentic Context Engineering (Evolving Playbook)
Each one covers how the technique works, when it earns its cost, and where it breaks.
The Agent Architect
One pattern, one tradeoff, one production failure story. A short weekly briefing for people building agentic systems.
Weekly email, one-click unsubscribe. We only use your address to send the briefing.