Caching actions after agent takes them #132
Description
After the LLM successfully completes an action would it be somehow possible to cache that action and call that action later in a subsequent run (even manually through playwright) without needing the LLM to take the "agentic" route to get there?