feat: ask_user (Q&A) + Monty run_python + SkillsToolset support

[OchnikBartek]

Summary

Adds three PydanticAI-only agent capabilities to the generated template, stacked on #83:

• ask_user — the agent pauses a run to ask the user one or more questions, then resumes with their answers (interactive multi-step card).
• run_python — a Monty-sandboxed code-execution tool: compute, transform data, and emit live charts/maps in a single turn.
• SkillsToolset — loads SKILL.md files from backend/skills/ as agent tools (loader only; drop in your own skills).

Changes

ask_user (Q&A)

• ask_user tool: agent pauses, WebSocket session pauses/resumes, frontend QuestionPrompt multi-step card collects answers
• Drop a stray ask_user_response frame (duplicate / post-reconnect) instead of surfacing a spurious "Empty message" error

Code execution (Monty)

• enable_code_execution flag gating a run_python tool backed by pydantic-monty
• In-sandbox create_chart / create_map / current_datetime emit live, persisted cards via emit_tool_event
• AgentSession streams in-sandbox tool calls as the same tool_call/tool_result event pair as direct calls

Skills (SkillsToolset)

• enable_skills flag + CLI --skills + interactive prompt
• assistant.py: SkillsToolset registration (PydanticAI, gated on backend/skills/ existing — no-ops when empty)
• pyproject: pydantic-ai-skills dep
• Frontend: load_skill / list_skills tool-call rendering

CI / tests

• New matrix configs pydantic_ai_code_execution and pydantic_ai_skills so both paths are linted + type-checked (previously uncovered)
• Run ruff via uvx in the generated-project tests (matches the post-gen hook; avoids a stray VIRTUAL_ENV breaking local runs)

Testing

• Matrix tests: generated pydantic_ai with --code-execution and with --skills (+ charts/AntV); compile, pass ruff + ty
• test_template_integration.py green (CrewAI ty skipped — known py3.14 limitation)
• Unit tests: test_config.py / test_cli.py / test_prompts.py — 227 passed
• Linting passes: make lint (generator + generated backend)
• Frontend build verified locally
• Generated projects still work: PydanticAI + code-execution + skills + AntV