All pull requests

Pull requests

• fix(claude-agent-sdk): keep the assistant model when result model_usage covers the session

  #4036·Harsh23Kashyap opened Oct 12, 2026Contributor·3 tasks done·

  ·
  1

• chore: release main

  autorelease: pending
  #4035·mikeldking opened Oct 12, 2026Contributor·

  ·

• fix(bedrock): JS - include cache tokens in Converse prompt token count

  #4032·chrikrah opened Oct 12, 2026·3 tasks·

  ·
  2

• fix(python): align GenAI dual-write and conformance harness with semantic-conventions-genai

  #4031·pardeep-singh opened Oct 12, 2026Contributor·

  ·
  1

• ci: exclude example projects from dependabot updates

  #4030·pardeep-singh opened Oct 12, 2026Contributor·

  ·
  1

• fix(ollama): end the span when a chat call is interrupted or cancelled

  #4029·mayuriphad opened Oct 12, 2026·

  ·
  1

  3

• chore(deps): update python-dotenv requirement from <2.0.0,>=1.2.3 to >=1.2.4,<2.0.0 in /python/instrumentation/openinference-instrumentation-pipecat

  dependenciespython
  #4028·dependabot[bot] opened Oct 12, 2026Bot·

  ·

• chore(deps): update anthropic requirement from <1.10 to <1.12 in /python/instrumentation/openinference-instrumentation-llama-index

  dependenciespython
  #4027·dependabot[bot] opened Oct 12, 2026Bot·

  ·

• chore(deps): update llama-index-llms-anthropic requirement from <0.13.0,>=0.12.1 to >=0.12.2,<0.13.0 in /python/instrumentation/openinference-instrumentation-llama-index

  dependenciespython
  #4026·dependabot[bot] opened Oct 12, 2026Bot·

  ·

• fix(openai): accumulate streamed refusal and reasoning content

  #4025·Garden12138 opened Oct 12, 2026·

  ·
  1

  2

• fix(openai-agents): set the root span to ERROR when the agent run fails

  #4024·Jennycruzy opened Oct 11, 2026·

  ·
  1

  9

• fix(openai): await AsyncAPIResponse.parse() for async .with_streaming_response calls

  #4020·darshjoshi opened Oct 11, 2026Contributor·3 tasks·

  ·
  1

  3

• feat(agno): trace external agents configured via Agno [WIP]

  #4018·kausmeows opened Oct 10, 2026Contributor·

  ·

• fix(litellm): record partial output on early stream close

  #4016·px12811 opened Oct 10, 2026·

  ·
  1

  2

• fix(litellm): set ERROR status on streaming spans that fail mid-stream

  #4015·chiliec opened Oct 10, 2026·

  ·
  1

  1

• fix(bedrock): record the system prompt and tools for Anthropic invoke_model

  #4014·pcbeingused333 opened Oct 10, 2026·

  ·
  1

  4

• fix(vercel): keep eve 0.76 agent.step and agent.approval spans as CHAIN

  #4007·jimbobbennett opened Oct 9, 2026Contributor·

  ·
  1

  1

• fix(agno): don't resolve an async tools factory in run attributes

  #3998·inchang-ing opened Oct 9, 2026·

  ·
  1

  3

• fix(openai): accumulate streamed refusal and reasoning_content

  #3996·HarshRajSinghania opened Oct 9, 2026·

  ·
  1

  2

• fix(google-genai): record tuple-valued contents as input messages

  #3995·inchang-ing opened Oct 9, 2026·

  ·
  1

  1

• fix(bedrock): record Converse model request fields in invocation parameters

  #3984·Harsh23Kashyap opened Oct 8, 2026Contributor·

  ·
  1

  1

• fix(bedrock): support file-like invoke_model body

  #3983·vineoy opened Oct 8, 2026·

  ·
  1

  5

• Fix #3904: do not record the SDK placeholder tool arguments on early stream exit

  #3982·inchang-ing opened Oct 8, 2026·

  ·
  1

  3

  

• Fix #3803: record attributes for tuple-valued and unsupported message content parts

  #3981·inchang-ing opened Oct 8, 2026·

  ·
  1

  3

• docs(instrumentation): add Sphinx API reference and Read the Docs config

  #3960·pardeep-singh opened Oct 6, 2026Contributor·3 tasks·

  ·
  1

  3