Issues

• after_model_callback: a replacement LlmResponse drops usage_metadata, erasing the model call from token accounting

  google/adk-python#7451 · Filter by author zhuhongd opened on Oct 7, 2026

• A2A executor returns raw exception details to remote peers

  google/adk-python#7445 · Filter by author alibaytok opened on Oct 7, 2026

• GKE code executor unit tests fail with kubernetes 37.0.0, turning main CI red

  google/adk-python#7443 · Filter by author vetler opened on Oct 7, 2026

• dev-ui 404s with url_prefix behind a prefix-stripping proxy since 2.11.0

  google/adk-python#7439 · Filter by author MihailRussu opened on Oct 7, 2026
  2

• LiteLlm streaming drops Gemini thought_signature from tool calls; Gemini 3 then rejects the follow-up request

  google/adk-python#7438 · Filter by author vetler opened on Oct 7, 2026
  2

• Legacy function declaration path raises NameError for a context param typed with a TYPE_CHECKING-only import

  google/adk-python#7436 · Filter by author Akshaay1 opened on Oct 7, 2026

• Tool confirmation: a repeated approval runs the confirmed tool again

  google/adk-python#7434 · Filter by author Vivek1106-04 opened on Oct 7, 2026
  1

  

• Update opentelemetry-api and opentelemetry-sdk to 1.44.0

  google/adk-python#7433 · Filter by author BrownMunda1 opened on Oct 7, 2026
  1

• A deterministic pre-execution layer for tool callbacks (before_back) — AgentDojo v2.2 re-run: ASR=0 / FP=0 (open artifacts)

  google/adk-python#7432 · Filter by author Lsy1533133 opened on Oct 7, 2026
  

• Parallel function calls: when one tool raises, finished siblings lose their responses and may be re-run

  Bug

  google/adk-python#7428 · Filter by author thuongvu opened on Oct 6, 2026
  1

  

• Ability to update instructions for PreLoadMemoryTool

  needs reviewtools
  Feature

  google/adk-python#7423 · Filter by author gmyrianthous opened on Oct 6, 2026
  1

  

• AgentEvaluator.evaluate() fails a lower-is-better metric whose evaluator reports PASSED

  eval
  Bug

  google/adk-python#7414 · Filter by author gaurav-gandhi-2411 opened on Oct 5, 2026
  2

  

• CI Mypy Check flags an existing streaming_utils.py error as new because the PR run reuses the baseline's mypy cache

  needs review
  Bug

  google/adk-python#7409 · Filter by author spandankeche opened on Oct 5, 2026
  1

  

• AgentTool runs the wrapped agent without the parent App's events_compaction_config, so agent-as-tool history grows unbounded

  Bug

  google/adk-python#7397 · Filter by author mahir-m01 opened on Oct 3, 2026
  5

  

• RemoteA2aAgent: streamed A2A chunks and the final event that replaces them get different event ids, so the answer can be shown twice

  a2a
  Bug

  google/adk-python#7395 · Filter by author ferponse opened on Oct 3, 2026
  2

  

• after_run_callback is not called when the task driving Runner.run_async is cancelled (regression in 2.11.0)

  core
  Bug

  google/adk-python#7394 · Filter by author roanny opened on Oct 3, 2026
  5

  

• A2aAgentExecutor sends the raw exception text to the A2A caller when the run fails

  a2arequest clarification
  Bug

  google/adk-python#7385 · Filter by author lamberttraccard opened on Oct 2, 2026
  2

  

• A task-mode root agent loses its task input once it asks the user a question

  core
  Bug

  google/adk-python#7384 · Filter by author lamberttraccard opened on Oct 2, 2026
  2

  

• FunctionTool fails to invoke callables with positional-only arguments

  tools
  Bug

  google/adk-python#7382 · Filter by author Moashehri opened on Oct 2, 2026
  7

  

• Please support mermaid 12 (inbuild elk) in adk web

  needs reviewweb
  Feature

  google/adk-python#7381 · Filter by author PhorstenkampFuzzy opened on Oct 2, 2026
  1

  

• rubric_based_final_response_quality_v1 errors when the agent replies from a before_agent_callback

  eval
  Bug

  google/adk-python#7379 · Filter by author MihailRussu opened on Oct 2, 2026
  1

  

• Feature request: Semaprax policy-plugin example for before_tool_callback

  needs reviewtools
  Feature

  google/adk-python#7378 · Filter by author wsdt opened on Oct 2, 2026
  2

  

• Allow LLM-as-judge evaluation metrics to use configured LiteLLM proxy credentials

  modelsneeds review
  Feature

  google/adk-python#7375 · Filter by author zmbrahian opened on Oct 2, 2026
  2

  

• LiteLlm drops GenerateContentConfig.thinking_config; map it to litellm's reasoning_effort

  models
  Bug

  google/adk-python#7360 · Filter by author chriswithers-fuse opened on Sep 30, 2026
  2

  

• Gemma model-family validation is skipped under python -O

  models
  Bug

  google/adk-python#7355 · Filter by author MateehUllah opened on Sep 30, 2026
  3