fix(runtime): preserve MCP tool inputSchema from SDK snake_case fields

[sdshah09]

Summary

• Fixed McpIntegration.list_tools() always returning empty inputSchema by reading the MCP SDK's snake_case input_schema attribute (camelCase is only the wire alias).
• Applied the same snake_case-first lookup for is_error / structured_content in _parse_result.
• Updated unit fakes and added a regression test so SDK-shaped attributes are exercised.

Fixes #1073

Test plan

• cd prime-agent-runtime && PYTHONPATH=src uv run python -m unittest discover -s test -p 'test_mcp_base.py' -v
• Confirmed _sdk_field returns the schema from a real mcp.types.Tool

Note

Fix MCP tool inputSchema population by reading SDK snake_case input_schema field

The MCP Python SDK exposes fields like input_schema, is_error, and structured_content as snake_case attributes, but the code was reading camelCase names (inputSchema, isError, structuredContent), causing tool schemas and structured results to be empty or ignored.

• Adds _sdk_field() helper in mcp_base.py that prefers the snake_case attribute and falls back to the camelCase alias, so both naming conventions are handled.
• Updates list_tools() to use _sdk_field(t, "input_schema", "inputSchema") so tool entries include the correct schema.
• Updates _parse_result() to read is_error and structured_content via the same helper.
• New tests verify snake_case fields are read correctly and camelCase aliases still work as fallbacks.

^{Macroscope summarized 33320d9.}

[sethkarten]

Thank you for the report and proposed work. This root cause is now covered by maintainer-owned stacked PR #1164, authored independently from upstream/main.

We did not inspect or reuse this PR's diff, branch, commits, implementation code, or tests; its public description/comments were used only as a bug report. To keep one review surface, this PR is superseded by #1164 and is being closed.

The complete review stack is #1158–#1165. It is being left unmerged for human review after CI and review-bot findings are cleared.