All pull requests

Pull requests

• test, docs: leftovers from the reviews of #43, #44, #59, #69 and #75

  #87·antonwolfy opened Oct 17, 2026Contributor·5 tasks done·

  ·
  1

• docs(benchmark): say when personas are discovered; pin the pinned run

  #86·antonwolfy opened Oct 17, 2026Contributor·5 tasks done·

  ·
  1

• feat(repo): --build-timeout for the image build of repo run and tasks check

  #85·antonwolfy opened Oct 17, 2026Contributor·5 tasks done·

  ·
  1

• fix(judge): --judge-provider alone picks a judge model that runs there

  #84·antonwolfy opened Oct 17, 2026Contributor·5 tasks done·

  ·
  1

• fix(tools): the cost cap sees a reply it cannot price; its calls are a prefix

  #83·antonwolfy opened Oct 17, 2026Contributor·5 tasks done·

  ·
  1

• docs: add VISION.md, a one-page statement of why dokimi exists

  #82·napetrov opened Oct 17, 2026Contributor·2 tasks done·

  ·
  7

• fix(judge): refuse a record with no question text on every judge path

  #81·napetrov opened Oct 17, 2026Contributor·5 tasks done·

  ·
  10

• refactor(eval): give answer and eval records one reader (eval/records.py)

  #80·napetrov opened Oct 16, 2026Contributor·5 tasks done·

  ·
  14

• feat(judge-validity): --context stored — gate the judge the way arms are scored

  #73·napetrov opened Oct 15, 2026Contributor·4 tasks done·

  ·
  39

• fix(novelty): a partial conflict check is not a cleared harmful_atoms pass

  #72·vchamarthi opened Oct 15, 2026Contributor·

  ·
  16

• feat(repo): repo budget, what a repository puts into an agent's context

  #71·napetrov opened Oct 15, 2026Contributor·4 tasks done·

  ·
  47

• fix(judge): read a qualified or negated figure verdict

  #67·napetrov opened Oct 15, 2026Contributor·4 tasks done·

  ·
  33

• repo inspect: static look at what a repository offers an agent (repo_environment.v1)

  #66·napetrov opened Oct 15, 2026Contributor·

  ·
  51

• docs: preparing a repository for coding agents

  #65·napetrov opened Oct 15, 2026Contributor·4 tasks done·

  ·
  40