Cursor rule Cursor
Cursor rule "cursorrules" from altaidevorg/rules-for-ai, covering agentic coding: humans design, agents code!, agentic coding steps, example llm project file structure, pocket flow and core abstraction.
Cursor rule
Flax NNX Filters (filterlib) for selecting Variables in Modules/State using types, tags, paths, predicates, and logical combinations.
Cursor rule
Explains Flax NNX GraphDef (static structure) and GraphState (dynamic values), the internal representation used by the functional API.
Cursor rule
Guidelines for using flax.
Cursor rule
Explains the Flax NNX Functional API (split, merge, state, update, graphdef) for bridging stateful Modules and JAX's functional paradigm.
Cursor rule
Details Flax NNX lifted transforms (nnx.jit, nnx.grad, nnx.vmap, nnx.scan) for applying JAX transformations directly to nnx.Module methods.
Cursor rule
Explains the core Flax NNX abstraction nnx.Module for defining neural network components as stateful Python classes.
Cursor rule
Explains Flax NNX nnx.Optimizer for managing Module, Optax state, and optimizer updates in training loops.
Cursor rule
Details Flax NNX nnx.Rngs for managing PRNG keys within Modules, explaining streams, key generation, and state management.
Cursor rule
Explains Flax NNX nnx.Variable (mutable state container in Modules) and nnx.VariableState (immutable JAX-compatible state).
Cursor rule
Tutorial chapter for google-adk's Agent (BaseAgent / LlmAgent), detailing the core agent abstraction, its structure, hierarchy, and the LLM-powered implementation.
Cursor rule
In the previous chapter, we learned how Event objects capture the history of interactions within a session. Many of these events, particularly agent responses and decisions to use tools, originate from interactions with a Large Language Model (LLM). This chapter introduces BaseLlm, the core abstraction in google-adk…
Cursor rule
Details google-adk's BaseLlmFlow, the core engine orchestrating LLM interactions, processors, tool calls, and agent transfers for LlmAgent.
Cursor rule
Details google-adk's Event object, representing interaction units with author, content, actions, and metadata within a Session.
Cursor rule
In the previous chapter, we explored BaseLlmFlow, the engine that orchestrates the complex interaction between an Agent (BaseAgent / LlmAgent) and the BaseLlm, especially when handling Tools (BaseTool). We saw that the flow logic often needs access to the current session, services, agent configuration, and more.…
Cursor rule
Tutorial chapter for google-adk's Runner, detailing its role in orchestrating agent execution, managing sessions, artifacts, memory, and selecting agents.
Cursor rule
Details google-adk's Session object for conversational history/state and BaseSessionService interface for session management.
Cursor rule
Details google-adk's State object for managing session state with scopes and delta tracking.
Cursor rule
In the previous chapter, we explored BaseLlm, the abstraction enabling agents to leverage the generative power of Large Language Models. However, LLMs alone have limitations – their knowledge is often frozen in time, and they cannot directly interact with the outside world or execute specific tasks beyond text…
Cursor rule
In the previous chapter, we learned about InvocationContext, the object encapsulating all necessary information for a single agent turn. While components like callbacks receive a CallbackContext that wraps InvocationContext and provides delta-aware state access, Tools (BaseTool) require slightly more specific…
Cursor rule
Explains google-genai API Modules (Models, Chats, Files, etc.) - facades grouping related API actions accessed via the Client.
Cursor rule
Details google-genai's APIError, the base exception for API service errors, parsing code, message, status from non-200 HTTP responses.
Cursor rule
Details the internal google-genai BaseApiClient, the core component for HTTP communication, auth, and endpoint handling.
Cursor rule
Explains google-genai Chat/AsyncChat classes, stateful wrappers for managing multi-turn conversation history with generative models.