v0.6.0

BREAKING: the default backend with no magi.toml/env changed from Anthropic
to Ollama (http://localhost:11434/v1, kimi-k2.6:cloud + the qwen3.5/gpt-oss/deepseek
trio). Anthropic still works but is now opt-in (provider = "anthropic" or
MAGI_PROVIDER=anthropic).

Added

• src/defaults.rs — single source of truth for the built-in default profile.
• --init-config CLI flag and /init-config TUI command to scaffold a default magi.toml.
• Startup notice (when no magi.toml) and an actionable error (when Ollama is unreachable),
  both DRY-interpolated from the default constants.

Changed

• resolve_provider/resolve_openai_base_url defaults → Ollama-first.
• resolve_openai_model no longer errors when unset — returns the built-in default.
• The MAGI trio defaults to qwen3.5/gpt-oss/deepseek on the openai path when [magi] is absent.

Known limitations

• The built-in defaults assume Ollama. If you point provider=openai at real OpenAI
  (or another non-Ollama service) WITHOUT setting OPENAI_MODEL / [openai].model / [magi],
  the defaults (kimi-k2.6:cloud + the :cloud trio) will not exist there — set them
  explicitly.
• The default :cloud model tags reflect the Ollama catalog at release time and may rot over
  time; refresh per release. They live in one place (src/defaults.rs) for easy maintenance.

Full Changelog: v0.5.2...v0.6.0