feat(mcp): add code navigation tools (code_outline / code_search / code_expand)

[wlff123]

Three new MCP tools backed by the existing @mcp.tool() endpoint, letting agents navigate code via structural outlines, leaf-name symbol search, and targeted symbol expansion — all operating on viking:// URIs only.

• New code_tools.py pure-formatter module turns CodeSkeleton into the three agent-facing text views.

• All 9 language extractors (Python, JS/TS, Go, Rust, Java, C/C++, C#, PHP, Lua) now populate line_start / line_end on every class, method, and function via tree-sitter's start_point / end_point (1-indexed). Skeleton dataclasses default these fields to 0 for back-compat.

• code_search caps at 200 files (warns the agent on overflow) and uses Semaphore(10) for fs.read concurrency, matching the existing read tool.

• Extension filter is applied before any fs.read so unsupported files (e.g. *.md) never hit storage.

• Full coverage: 18 pure-function tests in tests/parse/test_code_tools.py plus integration tests in tests/server/test_mcp_endpoint_code.py for URI validation, filtering, capping, and the three tool entry points.

Testing

• I have added tests that prove my fix is effective or that my feature works
• New and existing unit tests pass locally with my changes
• I have tested this on the following platforms:
  • Linux
  • macOS
  • Windows

Checklist

• My code follows the project's coding style
• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• My changes generate no new warnings
• Any dependent changes have been merged and published

Screenshots (if applicable)

Additional Notes

[github-actions]

PR Reviewer Guide 🔍

Here are some key observations to aid the review process:

⏱️ Estimated effort to review: 3 🔵🔵🔵⚪⚪

🏅 Score: 85

🧪 PR contains tests

🔒 No security concerns identified

✅ No TODO sections

🔀 Multiple PR themes Sub-PR theme: Add line numbers to AST skeletons Relevant files: openviking/parse/parsers/code/ast/skeleton.py openviking/parse/parsers/code/ast/extractor.py openviking/parse/parsers/code/ast/languages/*.py Sub-PR theme: Add code navigation MCP tools Relevant files: openviking/parse/parsers/code/ast/code_tools.py openviking/server/mcp_endpoint.py tests/parse/test_code_tools.py tests/server/test_mcp_endpoint_code.py docs/design/2026-05-20-code-tools-design.md

⚡ Recommended focus areas for review Minor log message regression The new extract() method's warning log no longer includes "falling back to LLM", which was present in the old extract_skeleton() path. Behavior is unchanged, but log observability is slightly reduced. def extract(self, file_name: str, content: str) -> Optional[CodeSkeleton]: """Return raw CodeSkeleton, or None for unsupported language / parse failure. Consumers that need the structured object (line numbers, traversal) call this; consumers that want the legacy formatted text use extract_skeleton. """ lang = self._detect_language(file_name) extractor = self._get_extractor(lang) if extractor is None: return None try: return extractor.extract(file_name, content) except Exception as e: logger.warning( "AST extraction failed for '%s' (language: %s): %s", file_name, lang, e ) return None

[github-actions]

PR Code Suggestions ✨

No code suggestions found for the PR.