Support standard YAML skill frontmatter

[richuV]

Summary

• replace the ad-hoc flat frontmatter parser with SnakeYAML-based parsing
• keep exposing only top-level scalar metadata values so the existing skill API stays compatible
• add regression tests for nested YAML fields and multiline scalar frontmatter

Why

MarkdownSkillParser currently rejects common YAML frontmatter constructs such as lists and block scalars. That makes otherwise valid SKILL.md files fail to load even when the framework only needs scalar fields like name and description.

This change makes the parser more tolerant of standard YAML frontmatter while preserving the current Map<String, String> metadata contract.

Testing

• mvn -pl agentscope-core "-Dtest=MarkdownSkillParserTest,SkillFileSystemHelperTest" test
• mvn -pl agentscope-core spotless:check

[cla-assistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[codecov]

Codecov Report

❌ Patch coverage is 71.42857% with 4 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
...gentscope/core/skill/util/MarkdownSkillParser.java	71.42%	1 Missing and 3 partials ⚠️

📢 Thoughts on this report? Let us know!

[jujn]

hi, you can take a look at this PR: #1034