fix: 🤖 attempt fix for issue #514

[github-actions]

🤖 Automated fix attempt

Fixes #514

Source issue: #514

Summary

Summary

Files changed:

1. mcp/src/tools/logs.ts — Added a direct searchClsLog MCP tool alias alongside the existing queryLogs tool:

   • Extracted the log search handler logic into a shared handleSearchClsLog() function
   • queryLogs(action="searchLogs") now delegates to handleSearchClsLog instead of duplicating the logic
   • Registered a new searchClsLog tool that calls handleSearchClsLog directly, with queryString as a required parameter (no action param needed)
   • Updated queryLogs description to note that action=searchLogs is equivalent to the searchClsLog tool, following the existing pattern in env.ts for legacy tool name documentation

2. mcp/src/tools/logs.test.ts — Added two test cases for the new alias:

   • searchClsLog tool should be registered and search cls logs — verifies tool registration, metadata, and that the handler calls the Manager SDK correctly
   • searchClsLog should error when queryString is missing — verifies the required-parameter validation

Design rationale: The searchClsLog tool exposes the canonical CloudBase Manager SDK method name as a direct MCP tool entry, so models/evaluators that already know the searchClsLog concept can discover and use it without translating through queryLogs(action="searchLogs"). Both tools share the same underlying handler, avoiding code duplication.

[binggg]

Follow-up iteration for issue_mnt9nu7f_yh77wx / run atomic-js-cloudbase-cli-log-query-by-request-id/2026-04-11T00-48-39-nszg86:

• clarified queryLogs(action="searchLogs") as the alias of searchClsLog
• returned explicit action/aliasOf metadata from both log search entries
• tightened the direct searchClsLog runtime guard for missing queryString
• refreshed focused vitest coverage for both entrypoints

Validation:

• npx vitest run mcp/src/tools/logs.test.ts
• npm --prefix mcp run build

The local AI Coding Eval Report API is currently unavailable on 127.0.0.1:5174 in this environment, so I could not run a fresh post-PR evaluation yet.