Google Antigravity SDK 0.1.11
Safe to upgrade? No issue tracker data
This tool has no public GitHub issue tracker we can query, so no signal is computed.
Changelog: Default Gemini 3.7 Flash upgrade, session budget limits, Vertex AI Express Mode, and autonomous behavior modes
The 0.1.11 release updates the default model to gemini-3.7-flash, introduces session-level budget enforcement and turn termination stop reasons, Vertex AI Express Mode authentication, and an autonomous agent behavior setting. It also expands tool hook metadata, resolves string annotation coercion for postponed evaluation, and improves MCP server and subagent stability.
Improvements(6)
- Default model upgrade to Gemini 3.7 Flash: Upgraded the default inference model to gemini-3.7-flash.
- Session budget enforcement and stop reasons: Added BudgetConfig to define session-level usage limits (total tokens, turns, and cost) and StopReason enum (BUDGET_EXCEEDED, TURN_LIMIT, USER_CANCELLED, etc.) to inspect turn termination causes.
- Vertex AI Express Mode support: Added native support for Express Mode authentication using VertexEndpoint(api_key=...) and LocalAgentConfig(vertex=true, api_key=...), simplifying headless and non-GCP deployments.
- Autonomous agent behavior mode: Control the agent's behavior with AgentBehavior. By default the SDK now has an AgentBehavior.AUTONOMOUS mode (previously AgentBehavior.INTERACTIVE) to streamline scripting, background, and headless interaction modes. Override by setting CapabilitiesConfig(agent_behavior=AgentBehavior.INTERACTIVE).
- Multi-interface hook registration: Enabled single-instance registration across multiple hook interfaces (PreToolHook, PostToolHook, PreTurnHook), allowing for cross-functional instrumentation without duplicate invocations.
- PreToolArgs metadata: Exposed trajectory_id and step_index on tool hook payloads for chat thread context tracing.
Fixes(4)
- ToolRunner string annotation coercion: When using from __future__ import annotations, tool argument coercion failed on stringified types; resolved by resolving type annotations using typing.get_type_hints before type adaptation.
- MCP server example port binding: Ephemeral port race conditions in test and example server startup were resolved by binding directly to port 0.
- Subagent deadlock prevention: Handled subagent fatal errors in the localagent executor to prevent deadlocks when subagents terminate abnormally.
- Empty input validation: Added input validation to prevent SDK unresponsiveness on empty or whitespace-only prompts.
<h3 class="