Expand description
Built-in tools — server-side tools, callable tools, and retrieval tools.
Re-exports§
pub use bash_tool::BashToolPolicy;pub use bash_tool::ExecuteBashTool;pub use example_tool::Example;pub use example_tool::ExampleProvider;pub use example_tool::ExampleTool;pub use exit_loop::ExitLoopTool;pub use get_user_choice::GetUserChoiceTool;pub use google_search::GoogleSearchTool;pub use load_memory::LoadMemoryTool;pub use long_running::LongRunningFunctionTool;pub use mcp::McpConnectionParams;pub use mcp::McpError;pub use mcp::McpSessionManager;pub use mcp::McpTool;pub use mcp::McpToolset;pub use preload_memory::PreloadMemoryTool;pub use retrieval::BaseRetrievalTool;pub use retrieval::FilesRetrievalTool;pub use retrieval::RetrievalResult;pub use retrieval::VertexAiRagRetrievalTool;pub use transfer_to_agent::TransferToAgentTool;pub use url_context::UrlContextTool;pub use vertex_ai_search::DiscoveryEngineSearchTool;pub use vertex_ai_search::VertexAiSearchConfig;pub use vertex_ai_search::VertexAiSearchTool;
Modules§
- bash_
tool - Bash execution tool — allows agents to execute shell commands.
- example_
tool - Example tool — adds few-shot examples to LLM requests.
- exit_
loop - Exit loop tool — allows agents to break out of loop execution.
- get_
user_ choice - Get user choice tool — presents options to the user for selection.
- google_
search - Built-in server-side Google Search tool.
- load_
memory - Load memory tool — allows agents to search their memory store.
- long_
running - Long-running function tool wrapper.
- mcp
- MCP (Model Context Protocol) toolset — connect to MCP servers and use their tools.
- preload_
memory - Preload memory tool — preloads memory into the LLM context.
- retrieval
- Retrieval tools — provide RAG (Retrieval-Augmented Generation) capabilities.
- transfer_
to_ agent - Transfer-to-agent tool — enables agent handoff in multi-agent systems.
- url_
context - URL context tool — built-in server-side URL context grounding.
- vertex_
ai_ search - Vertex AI Search tool — enterprise search via Vertex AI.