Labels

Single agent invocation: identity, configuration, reasoning/planning loop

Composition above single agent: workflow execution, handoffs, delegation across agents

Embedding requests and their results

Evaluation of GenAI runs: evaluator identity, runs, scores, verdicts

Single LLM call: model identity, request parameters, response and finish info, choices

Agent memory: store operations, record lifecycle, search, memory-layer taxonomy

Shape/content of input and output messages, parts, instructions, modalities, citations

Retrieval and vector databases: operations and returned documents/scores

Telemetry only for streamed responses: TTFT, TPOT, inter-token-latency, partial events

Token usage: input/output counts, breakdowns (cached, reasoning), per-modality, totals

Tool definitions and tool calls: declaration, execution, tool-call/result content shape

Something isn't working

Pull request review dashboard

Improvements or additions to documentation

This issue or pull request already exists

New feature or request

Good for newcomers

Extra attention is needed

This doesn't seem right

Conventions in mcp.* namespace: MCP transport, tool listing/approval, MCP-specific semantics

Message parts: shape and content of parts in messages

Issue/PR is waiting on a response from the author

Pull requests that update python:uv code

Further information is requested

Issue/PR has been inactive and will be closed soon

This will not be worked on