All pull requests

Pull requests

• fix(agy): resume by conversation ID and fail fast on auth errors

  #631·omkargaikwad23 opened Sep 26, 2026Collaborator·

  ·

• fix(agy): authenticate from the metadata server when no ADC file exists

  #630·omkargaikwad23 opened Sep 26, 2026Collaborator·

  ·
  3

• ci: add a weekly release pipeline with smoke gate and rollback

  #629·omkargaikwad23 opened Sep 24, 2026Collaborator·

  ·

• feat(scorers): award partial credit for goal_completion

  #626·omkargaikwad23 opened Sep 22, 2026Collaborator·

  ·
  4

• feat(mcp_readability): reuse unchanged findings so counts track tool changes

  #624·akangsha7 opened Sep 22, 2026Collaborator·

  ·

• feat(mcp_readability): decide what a run can reuse from its baseline

  #623·akangsha7 opened Sep 21, 2026Collaborator·

  ·

• feat(mcp_readability): record the judge's inputs alongside the tool surface

  #622·akangsha7 opened Sep 21, 2026Collaborator·

  ·

• feat(spanner): recognize real data type and sanitize multi-line quoted CSV fixtures

  #620·aseering opened Sep 19, 2026Collaborator·

  ·
  1

• feat(mcp_readability): read back the previous daily judgement for an endpoint

  #615·akangsha7 opened Sep 17, 2026Collaborator·

  ·

• ci(nl2sql): gate pull requests on a two-tier smoke check

  #611·omkargaikwad23 opened Sep 17, 2026Collaborator·

  ·
  2

• feat(mcp_readability): reuse unchanged findings so counts track tool changes

  #604·akangsha7 opened Sep 14, 2026Collaborator·

  ·

• chore(main): release 1.19.0

  autorelease: pending
  #598·release-please[bot] opened Sep 10, 2026Bot·

  ·

• feat(cortado): add multi-turn intermediate SQL execution, Set Match d…

  #582·erinlimbogoogle opened Aug 26, 2026Collaborator·

  ·
  2

• feat: Add Cloud Run job targets to rebuild the dataset quality cache

  #572·omkargaikwad23 opened Aug 13, 2026Collaborator·

  ·

• chore(viewer): remove dead code, hoist imports, trim Dataset Quality cards

  #560·omkargaikwad23 opened Aug 9, 2026Collaborator·6/6·

  ·

• refactor(viewer): collapse five copies of get_results_dir into paths.py

  #559·omkargaikwad23 opened Aug 9, 2026Collaborator·5/6·

  ·

• perf(viewer): reuse click handlers instead of rebuilding them per render

  #556·omkargaikwad23 opened Aug 9, 2026Collaborator·4/6·

  ·

• perf(viewer): share one mtime-keyed cache for the trends CSV and dir listing

  #555·omkargaikwad23 opened Aug 9, 2026Collaborator·3/6·

  ·
  2

• perf(viewer): send only plotted columns to the D3 charts

  #554·omkargaikwad23 opened Aug 9, 2026Collaborator·2/6·

  ·

• perf(viewer): cache List summaries instead of storing them in State

  #553·omkargaikwad23 opened Aug 9, 2026Collaborator·1/6·

  ·
  1

• Fix 5-tuple unpacking error in EvalServicer.Interact()

  #540·saurabh-net opened Jul 31, 2026Collaborator·

  ·
  2

• feat(scorers): support multiple python_scorer_* entries in run YAML

  #528·saurabh-net opened Jul 27, 2026Collaborator·

  ·
  1

• [WIP] fix(dea): support local boq evaluation via stubby

  #525·minzznguyen opened Jul 24, 2026Contributor·

  ·

• Add support for per-prompt execution timeout (prompt_timeout_seconds)

  #493·graceqi-g opened Jul 14, 2026Collaborator·

  ·
  30

• Add field to track time to first response and token usages

  #492·totoleon opened Jul 13, 2026Collaborator·

  ·
  2