Google Antigravity Python SDK — v0.1.17 Release Notes

[karmel]

[0.1.17] - 2026-09-14

This release introduces first-class conversation compaction controls via CompactionConfig, delta and forward-looking budget scopes for session resumption, tool output token truncation limits, and expanded arithmetic operations on UsageMetadata. It also delivers OS-level terminal sandboxing examples and automated tool wrapper reflection preservation.

🌟 Key Highlights

• Compaction Configuration: Adds CompactionConfig on AgentConfig to govern sliding-window conversation history compaction through an explicit token ceiling dial (token_threshold), deprecating CapabilitiesConfig.compaction_threshold.

  from google.antigravity import CompactionConfig, LocalAgentConfig
  
  config = LocalAgentConfig(compaction_config=CompactionConfig(token_threshold=50000))

• Forward-Looking Budget Scope: Adds BudgetScope.FORWARD_LOOKING to BudgetConfig to enforce model call and token budgets across newly resumed execution turns without counting previous historical usage.

  from google.antigravity import BudgetConfig, BudgetScope, LocalAgentConfig
  
  config = LocalAgentConfig(budget_config=BudgetConfig(max_total_tokens=10000, scope=BudgetScope.FORWARD_LOOKING))

• Tool Output Token Truncation: Exposes tool_output_truncation_config across agent configurations to cap token output volume initially for run_command executions in localharness.

  from google.antigravity import CapabilitiesConfig, LocalAgentConfig, ToolOutputTruncationConfig
  
  config = LocalAgentConfig(
      capabilities=CapabilitiesConfig(
          tool_output_truncation_config=ToolOutputTruncationConfig(max_tokens=2048)
      )
  )

• UsageMetadata Arithmetic: Supports standard Python arithmetic protocols on UsageMetadata, enabling scalar multiplications, scaling, and accumulating token counts using built-in sum().

  from google.antigravity.types import UsageMetadata
  
  total_usage = sum([usage1, usage2], start=UsageMetadata())
  scaled_usage = total_usage * 1.5

---

📋 Detailed Changes

Features & Enhancements

• Tool Wrapper Metadata Preservation: Preserves function signature, module name, annotations, and original callable references via __wrapped__ when registering tools with ToolWithSchema.
• Interactive REPL Policy Flattening: Automatically flattens nested policy lists during interactive REPL upgrades so nested command authorization rules correctly upgrade to prompt the user.
• Terminal Command Sandboxing Guide: Added getting-started guide and references demonstrating OS-level command sandboxing using RunCommandConfig(enable_sandbox=True) paired with execution policies.

Bug Fixes

• Tool Call Deserialization: Fixed dropped tool arguments during tool call handling when incoming payloads provide structured dictionary arguments rather than serialized JSON strings.
• Local Step Trajectory Tracking: Fixed missing provenance metadata by forwarding trajectory_id from incoming tool calls to local connection execution steps.
• Dynamic Content Proto Resolution: Fixed an AttributeError when importing struct_converter in external environments missing internal protobuf definitions by resolving descriptor types dynamically at runtime.