Learning Agent Optimizer FAQs

Question 1

Which AI model is used for the learning process?

Accepted Answer

The skill is configured to use the Sonnet model for task spawns to provide an optimal balance between reasoning quality for root cause analysis and operational cost.

Question 2

How does it handle multiple agent sessions at once?

Accepted Answer

It uses a hybrid approach: it identifies issues in parallel across all sessions, then processes investigation and updates serially for sessions from the same agent to prevent conflicts, while running different agents in parallel.

Question 3

Does this skill modify my agent files directly?

Accepted Answer

No, the skill follows a strict guardrail of delegating file modifications to specialized learning-cycle tasks and the learning-agent-expert to ensure changes are accurate and controlled.

Question 4

What is the primary function of the learn skill?

Accepted Answer

The learn skill processes pending LearningAgent session transcripts to identify performance issues, investigate their root causes, and update agent definitions to improve future behavior.

Question 5

What happens if a specific session fails to process?

Accepted Answer

If a sub-task fails, the skill logs the failure, skips that specific session, and continues with the remaining queue. The failed session remains marked as pending for future review.

Learning Agent Optimizer

Key Features

Use Cases

Learning Agent Optimizer

Key Features

Use Cases