docs(apify-actor-development): clarify AI / LLM Actors are just Actors that call an LLM

[DaveHanns]

Rationale

Users building LLM-powered Actors sometimes reach for agent frameworks or assume there is a special "AI Actor" construct in Apify. There isn't — an AI Actor is just an Actor whose work happens to be calling an external LLM API, with the same packaging conventions as any other Actor. A one-paragraph framing under apify-actor-development clarifies the mental model up front and steers builders away from over-engineering.

Change

Adds a short standalone ## AI / LLM Actors section between ## Best practices and ## Logging in skills/apify-actor-development/SKILL.md.

Added content (verbatim)

AI / LLM Actors

An "AI Actor" is just an Actor whose work happens to be calling an external LLM API — not a distinct construct, not an agent framework. Use the model vendor's SDK (e.g. @anthropic-ai/sdk, openai) to call the model, and the Apify SDK for input parsing, dataset writes, key-value store, and proxy config. Package it like any other Actor.

Surfaced during an evaluation of Apify surfaces for agent-driven Actor development.