AGOPSstill growing

Evaluating and running agents

Test an agent before release, watch it after, and follow the lifecycle that connects the two.

For Developers who can build an agent and now need to know whether it is good and whether it stays good.

PrerequisitesAGENTS Agents from first principlesAGSAFE Safety and guardrails for agents

7 notes, in order. Each one assumes the one before it. Still growing: more notes may be added.

Notes in this trail

Start the trail
  1. I · Evaluating
    Agent evaluation, not yet read
  2. Test datasets and mock tools, not yet read
  3. LLM as judge, not yet read
  4. II · Running
    Agent ops, not yet read
  5. Traceability, not yet read
  6. Agent metrics, not yet read
  7. III · The lifecycle
    The agent development lifecycle, not yet read