refactor: refactor chat service

[bytecii]

Description

refactor 1 / n

Source split:

backend/app/service/chat_service.py
  → backend/app/service/chat_service/

├── __init__.py              — Public API re-exports
├── step_solve.py            — Main async generator loop (was step_solve function)
├── question_router.py       — Query classification + simple answer routing (was question_confirm/simple_answer calls + routing logic)
├── handlers.py              — Action handlers: improve, pause, resume, end, new_task, etc.
├── lifecycle.py             — Agent/workforce construction: construct_workforce, new_agent_model, install_mcp
├── decomposition.py         — Task decomposition: make_sub_tasks, start_workforce
└── types.py                 — StepSolveState dataclass (shared mutable state for the loop)

Functions moved to agent layer:

chat_service.py → backend/app/agent/factory/
├── question_confirm.py      — question_confirm() + simple_answer() + _create_question_agent()
│                              (were inline in chat_service, now cached on task_lock.question_agent)
├── task_summary.py          — summary_task() + summary_subtasks_result() + get_task_result_with_optional_summary()
│                              (inference functions moved to agent layer where they belong)
└── workforce_agents.py      — create_coordinator_and_task_agents() + create_new_worker_agent()
                               (were inline in construct_workforce, extracted as factory functions)

Shared constants extracted:

chat_service.py → backend/app/agent/prompt.py

├── AGENT_ENVIRONMENT_PROMPT          — Platform/working-dir/date template (was duplicated 4x inline)
├── DEVELOPER_WORKER_DESCRIPTION      — Workforce agent description (was inline string in construct_workforce)
├── BROWSER_WORKER_DESCRIPTION        — Same
├── DOCUMENT_WORKER_DESCRIPTION       — Same
└── MULTI_MODAL_WORKER_DESCRIPTION    — Same

Context utilities extracted:

chat_service.py → backend/app/utils/context.py

├── build_conversation_context()      — Formats task_lock.conversation_history
├── collect_previous_task_context()   — Gathers previous task content + generated files
└── format_task_context()             — Formats a single task context block

Testing Evidence (REQUIRED)

• I have included human-verified testing evidence in this PR.
• This PR includes frontend/UI changes, and I attached screenshot(s) or screen recording(s).
• No frontend/UI changes in this PR.

What is the purpose of this pull request?

• Bug fix
• New Feature
• Documentation update
• Other

Contribution Guidelines Acknowledgement

• I have read and agree to the Eigent Contribution Guideline