fix: MCP queryFunctions output 结构嵌套导致 AI 解析出错

[binggg]

Attribution issue

• issueId: issue_moisa1j5_a1ora6
• category: tool
• canonicalTitle: MCP queryFunctions output 结构嵌套导致 AI 解析出错
• representativeRun: atomic-js-cloudbase-list-functions/2026-04-28T15-28-11-663142

Automation summary

• root_cause: The queryFunctions tool's listFunctions action returned function objects using Tencent Cloud API format with FunctionName field (capital letters). The AI had difficulty parsing this nested structure, leading to incorrect function name extraction and listing non-existent functions in internal evaluation artifact.
• changes: Modified mcp/src/tools/functions.ts to add a name field to each function in the listFunctions output. The fix maps each function object to include name: fn.FunctionName while preserving all original fields. This makes it easier for AI to extract function names using the intuitive name field.
• validation:
• TypeScript compilation passes without errors
• Skill/doc tests pass (10/10 tests passed)
• follow_up: None. This is a minimal fix addressing the root cause - the output now includes both name (easy to parse) and the original FunctionName (for backward compatibility).

Changed files

• mcp/src/tools/functions.ts

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.

[binggg]

Attribution post-PR evaluation

• visibility: internal identifiers, run ids, and private links are intentionally omitted
• attempt: 1
• eval_scope: primary_only
• overall: ERROR
• summary: post-PR evaluation errored
• updated_at: 2026-04-28T16:47:06.285Z

Cases

• [ERROR] — primary — evaluation errored