Production
Observability, security, evals, CLI — everything you need to trust an agent in production.
Shipping an agent is not just about getting the answer right once. It is about making the system observable, safe, debuggable, and repeatable under real traffic.
Start with the shipping checklist if you want the shortest path from prototype to rollout.
#Rollout
Move from prototype to trustworthy system.
#Observability
Trace, log, audit, cost-guard every run.
- Overview · Loggers · Trace viewer · Devtools · Cost guard · Audit log
#Security
Six primitives every production agent needs.
#Evaluation
Measure quality with numbers, not vibes.
- Overview · Suites · Replay · Snapshots · CI reporters
#CLI
Nine commands wrapping every part of AgentsKit.
Explore nearby
- PeerShipping checklist
A practical checklist for taking an AgentsKit agent from prototype to production.
- PeerOn-call runbooks
First-response playbooks for the four most common AgentsKit production incidents — LLM provider outage, tool flapping, cost spike, prompt injection.
- PeerPerformance budgets
Bundle size ceilings per package, enforced in CI via size-limit. Measured values injected when available.