All pull requests

Pull requests

• feat(compare): compare multiple offline runs with HTML and Markdown reports

  #353·zpzjzj opened Oct 29, 2026Collaborator·4 of 5 tasks·

  ·
  1

  1

• feat(evaluator): expose benchmark variant to setup scripts

  #352·zpzjzj opened Oct 29, 2026Collaborator·

  ·
  1

  1

• feat(judge): validated named metrics and persistence (script judge, slice 1)

  #347·GBX-Max1220 opened Oct 27, 2026·

  ·
  7

• fix(config): reject empty and incomplete rule_based assertions in validate

  #341·hobostay opened Oct 26, 2026·2 tasks done·

  ·
  1

• chore: bump @deepseek-ai/cordis from 4.0.2 to 4.0.5 in /plugins/dsh-skill-up in the deepseek-harness group across 1 directory

  #335·dependabot[bot] opened Oct 22, 2026Bot·

  ·
  4

• fix(codex): default custom provider wire_api to 'responses'

  #293·CZH-THU opened Oct 1, 2026·

  ·
  3

• feat(cli): check skill content integrity in validate

  #281·BiBoyang opened Sep 25, 2026Contributor·4 tasks done·

  ·
  5

• fix(qodercli): keep child tools on the evaluation session

  #219·chiga0 opened Aug 25, 2026Contributor·

  ·
  6

• fix(qoder): match underscore workspace keys

  #170·chiga0 opened Aug 4, 2026Contributor·

  ·
  3

• feat(runtime): support Windows OpenSandbox guests

  #148·zpzjzj opened Jul 17, 2026Collaborator·

  ·
  1

  9