proactive-self-improving-agent

Captures useful lessons from errors, corrections, and better practices, then promotes recurring lessons into lasting agent rules.

When to use it

Use when work fails, the user corrects the agent, a capability is missing, knowledge is outdated, a better practice emerges, or a task finishes.

On seven trigger scenarios, it evaluates novelty and writes the relevant learning, error, or feature record into the project.