Issues

• Confirming the tool calls of two parallel LlmAgent nodes in one message runs only one tool; the other agent's model gets "This tool call requires confirmation"

  google/adk-python#7685 · Filter by author yifanxiong272 opened on Oct 22, 2026

• A Workflow LlmAgent node with two pending tool confirmations does not keep waiting when only one is answered; the next node runs too early

  google/adk-python#7684 · Filter by author yifanxiong272 opened on Oct 22, 2026
  1

• ToolConfirmation.from_response_dict crashes with TypeError on parsed dicts

  google/adk-python#7681 · Filter by author aryanpatil-07 opened on Oct 22, 2026

• Retrieval tools declare query as optional, and a call without it fails the run

  Bug

  google/adk-python#7677 · Filter by author roydonsequeira opened on Oct 22, 2026
  

• LlamaIndexRetrieval and FilesRetrieval block the event loop on every lookup

  Bug

  google/adk-python#7674 · Filter by author roydonsequeira opened on Oct 22, 2026
  

• Gemma 3 models fail with TypeError when a tool returns a datetime or a Pydantic model (e.g. load_memory)

  models
  Bug

  google/adk-python#7667 · Filter by author roydonsequeira opened on Oct 22, 2026
  6

  

• Workflow: a node that returns None runs again when the workflow resumes after RequestInput

  Bug

  google/adk-python#7666 · Filter by author AkshayShah03 opened on Oct 22, 2026
  1

  

• AnthropicLlm sends failed MCP tool results to Claude with is_error=False

  mcpmodels
  Bug

  google/adk-python#7664 · Filter by author AkshayShah03 opened on Oct 22, 2026
  

• FirestoreSessionService.list_sessions(user_id="") returns every user's sessions

  services
  Bug

  google/adk-python#7662 · Filter by author AkshayShah03 opened on Oct 22, 2026
  2

  

• Parallel tool calls: a dict state key removed by a later call comes back in the persisted state

  tools
  Bug

  google/adk-python#7661 · Filter by author AkshayShah03 opened on Oct 22, 2026
  

• AnthropicLlm leaves error_code empty when Claude refuses or hits max_tokens with no content

  models
  Bug

  google/adk-python#7660 · Filter by author chrikrah opened on Oct 22, 2026
  

• Duplicate rubric verdicts let one judge sample outvote two independent samples

  eval
  Bug

  google/adk-python#7659 · Filter by author BichengWang opened on Oct 21, 2026
  4

  

• BigQueryAgentAnalyticsPlugin: log token usage for live (run_live) sessions

  bq
  Feature

  google/adk-python#7657 · Filter by author jahel-ax opened on Oct 21, 2026
  

• delete_files confirmation is a model-controlled tool argument

  tools
  Bug

  google/adk-python#7655 · Filter by author sai5880 opened on Oct 21, 2026
  1

  

• ApigeeLlm: streamed responses drop reasoning tokens and cached prompt tokens from usage_metadata

  models
  Bug

  google/adk-python#7654 · Filter by author pcbeingused333 opened on Oct 21, 2026
  1

  

• After a tool confirmation resumes a Workflow, an LlmAgent node's model receives the workflow's user message instead of the node's input

  Bug

  google/adk-python#7650 · Filter by author yifanxiong272 opened on Oct 21, 2026
  1

  

• Evaluation judge and user simulator default to gemini-2.5-flash, which is no longer available to new users

  evalmodels
  Bug

  google/adk-python#7649 · Filter by author pcbeingused333 opened on Oct 21, 2026
  

• ApigeeLlm raises ValueError instead of reporting MAX_TOKENS when streamed tool arguments are cut off

  tools
  Bug

  google/adk-python#7648 · Filter by author BichengWang opened on Oct 21, 2026
  3

  

• MCP invocation falls through empty OAuth2 token to HTTP field, diverging from discovery

  mcp
  Bug

  google/adk-python#7646 · Filter by author CorgiBoyG opened on Oct 21, 2026
  2

  

• ModelArmorPlugin fails with "Event loop is closed" after one sync Runner.run() call: its client is not cached per event loop

  agent engine
  Bug

  google/adk-python#7644 · Filter by author qingvincentyin opened on Oct 21, 2026
  1

  

• RFC: supervised task handles, queued steering, and explicit result merges

  core
  Feature

  google/adk-python#7642 · Filter by author me-her opened on Oct 21, 2026
  

• Support max_parallel_workers on Agent parallel workers

  agent config
  Feature

  google/adk-python#7641 · Filter by author me-her opened on Oct 21, 2026
  

• ReflectAndRetryToolPlugin and ReflectAndRetryModelPlugin keep failure counts for every finished invocation

  Bug

  google/adk-python#7637 · Filter by author Somay-kousis opened on Oct 20, 2026
  

• FunctionTool passes a plain str to an Enum-typed parameter, so the tool crashes

  tools
  Bug

  google/adk-python#7635 · Filter by author Somay-kousis opened on Oct 20, 2026
  4

  

• AnthropicLlm ignores tool_config: function calling mode NONE or ANY never reaches Claude

  models
  Bug

  google/adk-python#7633 · Filter by author Somay-kousis opened on Oct 20, 2026
  1