Agent Introspection

The more I use my clawdbot the more it seems important to have the ability for an agent to introspect itself. What I mean is, it's very important for an agent to know exactly how its harness works, how to edit it to improve it, and basically also how to debug itself.

An agent should be able to answer truthfully and know its limitations, and not waffle and bullshit about capabilities it doesn't have.