Commit Graph

  • ba83d0cfe4 feat: LLM-first architecture — agent orchestrates everything via solve_task tool main Glevel 2026-06-11 18:43:40 +03:00
  • 0a2b139f9b fix: SyntaxError nonlocal — заменить на dict _pl_state Glevel 2026-06-05 17:32:45 +03:00
  • 101671ea6e feat: логи tool-calls для pipeline всех заданий Glevel 2026-06-05 17:31:05 +03:00
  • 2bfcf5f782 revert: вернуть ui.py к версии ee2c8de (до изменений rate limit) Glevel 2026-06-05 17:16:14 +03:00
  • b9b0f58de2 fix: переписать UI — убрать polling/threading, простой blocking spinner Glevel 2026-06-05 17:12:19 +03:00
  • 177dea2769 fix: переписать UI на st.rerun() polling — кнопки и события работают корректно Glevel 2026-06-05 17:05:09 +03:00
  • cba8940005 fix: 429 rate limit виден в UI, добавлена кнопка Стоп Glevel 2026-06-05 16:59:10 +03:00
  • ee2c8de372 fix: исправить поиск tasks_list — использовать полное имя с префиксом MCP Glevel 2026-06-05 16:49:48 +03:00
  • 3ae3488691 refactor: добавить агентную архитектуру, UI и CLI Glevel 2026-06-05 16:43:00 +03:00
  • bb1aedfe2a fix: read teacher feedback from submission.grade.feedback Glevel 2026-06-04 20:22:23 +03:00
  • da40caa637 feat: trap detection - validate teacher comment before rework Glevel 2026-06-04 19:52:36 +03:00
  • 7faae42204 feat: stronger defense arguments in rework analysis Glevel 2026-06-04 19:41:17 +03:00
  • 02ed3aa99f fix: detect rework via Gitea repo existence, not BroJS answer field Glevel 2026-06-04 19:30:30 +03:00
  • 39019e2574 feat: analyze teacher comments before rework (defend solution) Glevel 2026-06-04 19:21:39 +03:00
  • 07d7d01996 fix: ban Ollama in prompt, use OpenAIEmbeddings for ChromaDB RAG Glevel 2026-06-04 19:18:43 +03:00
  • ec197b7e98 docs: explain OpenRouter vs Ollama choice Glevel 2026-06-04 19:13:54 +03:00
  • 863ea54eaf fix: безопасная перекодировка stdout/stderr и рекурсивная проверка 429 Glevel 2026-06-04 19:00:14 +03:00
  • 9cb3731dda fix tasks command: show all tasks with statuses, not only todo Glevel 2026-06-04 13:52:34 +03:00
  • 943609f370 add console.py: interactive CLI for managing BroJS tasks Glevel 2026-06-04 13:50:47 +03:00
  • 65fe475e42 add rework detection to solve_task: read teacher comments, pass to LLM Glevel 2026-06-04 12:49:59 +03:00
  • 928dabe10e integrate fast solver into pipeline: run_pipeline.py now fully automated Glevel 2026-06-04 12:44:07 +03:00
  • bd7580c4bc optimize solve_task.py: persistent MCP client, exponential backoff, inter-call delay Glevel 2026-06-04 12:38:23 +03:00
  • 420543c89e optimize mcp_client: persistent client, exponential backoff, inter-call delay Glevel 2026-06-04 12:36:46 +03:00
  • 7804122b7e feat: add solve_task.py direct solver and fix MCP rate limit issues Glevel 2026-06-04 12:30:05 +03:00
  • 5f8b5b9144 add run/check scripts and revert llm to OpenRouter Glevel 2026-05-28 18:47:05 +03:00
  • 1d73c5e439 feat: pipeline guarantees task submission regardless of LLM tool calling Glevel 2026-05-26 22:57:24 +03:00
  • 7744a97012 chore: switch LLM from OpenRouter to BroJS Inference API Glevel 2026-05-26 18:58:41 +03:00
  • 5f4737ce12 fix: syntax error in prompts.py (triple quotes inside triple quotes) Glevel 2026-05-26 00:43:47 +03:00
  • 698ba21cfd improve: better prompts and rate-limit retry in pipeline Glevel 2026-05-26 00:35:02 +03:00
  • da0facfd4d fix: add __init__ to SanitizeToolCallsMiddleware Glevel 2026-05-25 23:27:25 +03:00
  • c95d7375cc feat: initial brojs-agent implementation Glevel 2026-05-25 23:08:14 +03:00