Add Markdown frontmatter support to the Visual Brief

[yoldaolmak]

Problem

The VisualBrief currently relies on headings for structure. Many articles carry standard YAML frontmatter (categories, tags, target audience) that should inform image selection context.

Scope

• src/pictova/core/primitives.py — where VisualBrief.from_markdown() parses the article.
• Parse a leading YAML frontmatter block if present; ignore cleanly if absent or malformed.
• Pass frontmatter fields into the brief so downstream vision context can use them.
• Unit tests: with frontmatter, without, and with malformed frontmatter.

Acceptance criteria

• Frontmatter variables are available on the brief and reach the vision context block.
• Articles without frontmatter behave exactly as before.

Difficulty: Low · Credentials: None.

[yoldaolmak]

Maintainer update: this remains a good first issue. The current parser lives in the core primitives module; preserve no-frontmatter behavior and add fixtures for valid and malformed frontmatter. Keep parsed fields explicit in the Visual Brief rather than introducing undocumented downstream state.