All pull requests

Pull requests

• 🤖 refactor: move token budget orchestration behind context management

  #4358·ThomasK33 opened Sep 23, 2026·3/3·

  ·
  1

• 🤖 fix: hide token-budget flush output from the transcript

  #4357·ThomasK33 opened Sep 23, 2026·2/3·

  ·
  1

• 🤖 tests: keep the full unit run green in one shared Bun process

  #4356·ThomasK33 opened Sep 23, 2026·1/3·

  ·
  1

• 🤖 tests: fix Pixel snapshot flakes and gate nondeterministic stories

  #4353·ammar-agent opened Sep 22, 2026·

  ·
  32

• 🤖 ci: shard critical-path PR jobs and dedupe workflow gates

  #4351·ammar-agent opened Sep 22, 2026·

  ·
  24

• 🤖 fix: send GPT-6 Sol/Luna reasoning effort none and repair main CI

  #4350·ammar-agent opened Sep 22, 2026·

  ·
  21

• 🤖 fix: price and kick off model-created goals on the invoking turn's model

  #4349·ammar-agent opened Sep 22, 2026·

  ·
  13

• 🤖 feat: let agents discover selectable models with models_list

  #4346·ThomasK33 opened Sep 22, 2026·

  ·
  7

• 🤖 fix: recover once from rejected OpenAI reasoning replay

  #4344·ThomasK33 opened Sep 22, 2026·2/2·

  ·
  10

• 🤖 fix: keep monitored sub-agents active in the composer

  #4341·ThomasK33 opened Sep 22, 2026·2/2·

  ·
  1

  3

• 🤖 fix(runtime): prevent non-regular plan files from starving filesystem workers

  #4339·ThomasK33 opened Sep 22, 2026·3/4·

  ·
  5

• 🤖 refactor: share workspace activity helpers

  #4338·ThomasK33 opened Sep 22, 2026·1/2·

  ·
  4

• 🤖 fix(history): recover excessively nested persisted tool payloads

  #4337·ThomasK33 opened Sep 22, 2026·2/4·

  ·
  4

• 🤖 fix(stream): reject excessively nested tool inputs before SDK parsing

  #4336·ThomasK33 opened Sep 22, 2026·1/4·

  ·
  4

• 🤖 fix: replay OpenAI reasoning without server-side item references

  #4335·ThomasK33 opened Sep 22, 2026·1/2·

  ·
  5

• 🤖 feat(workflows): evaluate() authoring docs and GitHub issue screening example

  #4333·ThomasK33 opened Sep 22, 2026·Queued·7/7
  41

• 🤖 feat(workflows): TypeSafe evaluation provider for evaluate()

  #4332·ThomasK33 opened Sep 22, 2026·Queued·6/7
  4

• 🤖 feat(workflows): evaluate() step rendering in the workflow timeline and tool card

  #4330·ThomasK33 opened Sep 22, 2026·Queued·5/7
  48

• 🤖 feat(workflows): evaluate() primitive wired through runner, service and CLI

  #4329·ThomasK33 opened Sep 22, 2026·Queued·4/7
  55

• fix(sidebar): continuous sub-agent connector lines

  #4325·jaaydenh opened Sep 22, 2026·

  ·

• 🤖 feat(workflows): evaluation adapter, step admission schema and evaluation model default (L3a)

  #4324·ThomasK33 opened Sep 22, 2026·Queued·3/7
  27

• 🤖 feat(evaluation): evaluation model resolver and Effect EvaluationService

  #4322·ThomasK33 opened Sep 21, 2026·Queued·2/7
  36

• 🤖 feat(evaluation): shared evaluation schemas, limits and answer validator

  #4321·ThomasK33 opened Sep 21, 2026·Queued·1/7
  25

• 🤖 chore(deps): pin ai 7.0.107 and @ai-sdk openai/anthropic/google for experimental_evaluate

  #4319·ThomasK33 opened Sep 21, 2026·

  ·
  5

• 🤖 feat: persist native plan review in workspace history

  #4317·ThomasK33 opened Sep 21, 2026·4/4·

  ·
  100