huzhenyu

huzhenyu слил пул реквест huzhenyu/agent-lab#1

init

1 неделя назад

huzhenyu закрыл запрос на слияние huzhenyu/agent-lab#1

init

1 неделя назад

huzhenyu создал запрос на слияние huzhenyu/agent-lab#1

init

1 неделя назад

huzhenyu запушил(а) develop в huzhenyu/agent-lab

538955a811 Initial commit

1 неделя назад

huzhenyu создал новую ветку develop в huzhenyu/agent-lab

1 неделя назад

huzhenyu запушил(а) codex/agent-lab-mvp в huzhenyu/agent-lab

63481a9a4c docs: design balanced ChatAgent event selection Problem: ChatAgent can verbally promise supported actions without emitting the text event required to execute them.\n\nRisk: stronger selection rules must preserve negative boundaries, legacy EventPrompt cleanup, and invocation-mode isolation.
32c0f1357e docs: record UI workflow observations Problem: the completed task needed durable notes for subagent Git fallback and explicit browser-QA waivers without expanding the core workflow.
df0e547c11 docs: record professional console merge Problem: the Todo 53 implementation plan still showed local merge as pending after the verified fast-forward integration.
8bec5bc095 docs: close professional console todo Problem: Todo 53 needed final verification evidence and an explicit record that browser QA was waived by the user.\n\nRisk: no browser result is represented as passing; later visual defects must use a new follow-up todo.
a10beb4eb4 docs: record console verification state Problem: the professional console needs durable operator guidance, completed follow-up evidence, and an explicit browser-QA blocker.\n\nRisk: Todo 53 remains open and must not be merged until desktop and narrow-screen visual checks pass.
Просмотр сравнение для этих 21 коммитов »

1 неделя назад

huzhenyu запушил(а) codex/agent-lab-mvp в huzhenyu/agent-lab

2f110357bf docs: close benchmark runner todo Problem: Todo 52 needed final evidence, completed follow-up status, operator-facing limitations, and workflow-evolution findings after implementation and repeated reviews. Risk: No live provider benchmark was executed; the closeout explicitly limits conclusions to deterministic mock/runtime behavior until the operator supplies a real target and credential.
81156abdbd fix: preserve provider stream exceptions Problem: TimingChatClient cleanup could replace an active provider failure or cancellation with an iterator close error. Risk: Close failures are now secondary only while a provider exception is propagating; normal completion and consumer-close paths continue to surface close failures.
16f3e2f9c1 docs: plan provider cleanup exception fix Problem: Iterator cleanup can replace an active provider error or cancellation, hiding root cause and breaking interruption semantics. Risk: Cleanup failures become secondary when another exception is active; tests must still prove standalone close failures propagate.
8e62d3c533 fix: preserve benchmark metric integrity Problem: model elapsed included consumer backpressure, missing provider usage produced false zero or partial token totals, and the parallel benchmark only checked successful event cardinality instead of handler overlap. Risk: providers that omit usage now produce failed benchmark samples with null token metrics; a non-overlapping parallel case may wait up to the 50 ms probe barrier timeout before completing.
e809f7fe92 docs: plan benchmark metric integrity fixes Problem: Model elapsed includes consumer backpressure, missing provider usage appears as zero tokens, and the parallel benchmark does not prove handler overlap. Risk: Adding a deterministic parallel probe and stricter usage completeness can turn previously passing live runs into explicit semantic failures when providers omit usage or execution becomes serial.

2 недель назад

huzhenyu создал новую ветку codex/agent-lab-mvp в huzhenyu/agent-lab

2 недель назад

huzhenyu создал(а) репозиторий huzhenyu/agent-lab

3 недель назад