A unified command definition framework for Go that generates CLI, HTTP, and MCP interfaces from a single definition
xyz-go exhibits significant quality gaps across naming, descriptions, and schema completeness. Tool names lack clear action verbs (e.g., 'user.add' should be 'create_user'; 'user.rm' should be 'delete_user'). Descriptions are minimal (10-20 chars), well below the 50-200 char LLM-optimized baseline. Parameter descriptions exist but are sparse. Schemas are present but lack depth, no min/max bounds, no format constraints, no output schema documentation. The 'net.head' tool exposes API keys as parameters (critical security violation). Error handling and recovery guidance are absent. Only 2 of 11 tools have descriptions exceeding 20 characters.
计算 SHA-256
两数相除
两数求和
探测 API Key 注入
搜索文档
监听端口
睡眠指定时长
当前 UTC 时间
API keys exposed as tool parameters (net.head). Credentials must never appear in parameters, use server-side secret injection via environment variables or vault. Agent traces log every parameter, leaking secrets into logs and prompt history.
All tool descriptions are under 20 characters (e.g., '创建用户', '列出用户', '搜索文档'). LLM-optimized descriptions should be 50-200 chars, explaining WHAT the tool does, WHEN to use it, and any prerequisites. Current descriptions are too terse for LLM selection.
Tool names lack clear action verbs. 'user.add' should be 'create_user', 'user.rm' should be 'delete_user', 'user.list' should be 'list_users'. Verb_noun naming lets LLMs parse intent from the name alone and disambiguate overlapping tools.
Inferred effective spec: 2025-06-18+.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 40 | 2025-06-18+ | v2 |
创建用户
列出用户
删除用户
No output schemas documented for any tool. LLMs need to know what fields to expect so they can plan downstream tool calls and extract the right data. Without output schemas, agents cannot chain tools reliably.
No error handling or recovery guidance. Tools do not indicate what to do on failure (retryable, user-fixable, fatal). Agents receive no actionable error messages to guide next steps.