AgentSkillsCN

performance-analyzer

分析查询与代码变更对性能的影响。能够识别 N+1 查询、缺失索引、同步循环中的性能瓶颈、未分页的 fetch 操作、缺乏缓存优化、低效算法等问题。适用于数据库查询变更、循环逻辑调整、API 接口更新、大型文件操作，或涉及 fetch/查询/循环/映射/过滤等关键字时使用。

SKILL.md
--- frontmatter
name: performance-analyzer
description: Analyzes performance impact of query and code changes. Detects N+1 queries, missing indexes, sync-in-loops, unpaginated fetches, missing memoization, inefficient algorithms. Use for database query changes, loops, API changes, large file ops, or keywords fetch/query/loop/map/filter.

Performance Impact Analyzer

When to Trigger

  • •Database query changes
  • •Adding loops or iterations
  • •API route modifications
  • •Large file operations
  • •Keywords: "fetch", "query", "loop", "map", "filter"

What to Do

Step 1: Detect Anti-Patterns

Check for:

  1. •N+1 queries
  2. •Missing database indexes
  3. •Synchronous operations in loops
  4. •Large data fetches without pagination
  5. •Missing memoization (React)
  6. •Inefficient algorithms (O(n²) or worse)

Step 2: Estimate Impact

Report current vs after (e.g. query count, time). Explain why slower (e.g. N+1).

Step 3: Suggest Optimizations

  • •Replace N+1 with single query using include/select or JOINs.
  • •Add indexes for filtered/sorted columns.
  • •Add pagination (skip/take or cursor).
  • •Use useMemo/useCallback where appropriate.
  • •Prefer single aggregation over multiple sequential queries.

Show SLOW vs FAST code snippets. Ask: "Apply optimized version? (yes/no)".