Issue #1055: Add minimal frontmatter parser for memory files

[Meh10t]

This PR starts work on issue #1055.

It introduces a minimal frontmatter parser for memory files, supporting arbitrary YAML-like fields.

Currently supported:

• Optional YAML-like frontmatter block at the top of a memory file
• Supports arbitrary frontmatter fields (no restriction to specific keys)
• Returns both the parsed frontmatter and the remaining body content

Example supported format:

---

summary: Some short description

This is the memory content...

This is the first step towards supporting opt-in frontmatter tools (MemoryAddFrontmatterTool and MemoryGetFrontmatterTool).

Next steps:

• Integrate frontmatter stripping in memory loading
• Implement optional frontmatter tools
• Extend list_memories behavior when frontmatter tools are enabled

[MischaPanch]

This is a draft, right?

[Meh10t]

This is a draft, right?

Yes, this is a draft.

This PR only introduces the minimal frontmatter parser as the first step for issue #1055.
The next steps will integrate it into memory loading and implement the optional frontmatter tools.

[MischaPanch]

Pls keep in mind our specification. In particular, any kind of frontmatter is allowed. So

Only the summary field is supported for now

Is not correct

[Meh10t]

Pls keep in mind our specification. In particular, any kind of frontmatter is allowed. So

Only the summary field is supported for now

Is not correct

Thanks for the clarification!

I updated the parser so it now accepts any frontmatter key/value pairs instead of restricting it to "summary".

Let me know if this aligns better with the specification.

[opcode81]

Serena is an object-oriented codebase. Please consistently use OOP principles.

[Meh10t]

Serena is an object-oriented codebase. Please consistently use OOP principles.

Thanks for the feedback! I refactored the frontmatter parsing into an OOP-style FrontmatterParser while keeping parse_frontmatter as a backward-compatible wrapper.