Agent Skills › franklioxygen/agent-workflows › performance-review

performance-review

GitHub

用于对代码变更、PR或分支进行性能审查，评估算法复杂度、数据库查询、N+1风险及内存使用等。通过扫描信号和追踪热点路径，提供具体的优化建议与验证方案，支持可扩展性分析与延迟风险评估。

skills/performance-review/SKILL.md franklioxygen/agent-workflows

Trigger Scenarios

请求代码性能审查请求可扩展性分析请求性能剖析计划请求基准测试计划请求延迟风险评估

Install

npx skills add franklioxygen/agent-workflows --skill performance-review -g -y

More Options

Use without installing

npx skills use franklioxygen/agent-workflows@performance-review

指定 Agent (Claude Code)

npx skills add franklioxygen/agent-workflows --skill performance-review -a claude-code -g -y

安装 repo 全部 skill

npx skills add franklioxygen/agent-workflows --all -g -y

预览 repo 内 skill

npx skills add franklioxygen/agent-workflows --list

SKILL.md

Frontmatter

{
    "name": "performance-review",
    "description": "Perform focused performance reviews of code changes, pull requests, branches, or workspace diffs for algorithmic complexity, database query patterns, N+1 risks, pagination, caching, batching, memory use, synchronous I\/O, latency, and load behavior. Use when Codex is asked for a performance review, scalability review, profiling plan, benchmark plan, or latency-risk assessment."
}

Performance Review

Quick Start

Run scripts/performance_signal_scan.py <path> when a repository or changed-file path is available.
Read references/performance-review-checklist.md before finalizing findings.
Follow shared safety, validation, and scope rules in ../_shared/references/skill-operating-rules.md.
Focus on hot paths, data size, request rate, and operational cost.
Stay read-only unless the user explicitly asks for fixes.

Review Workflow

Establish expected scale: records, requests per second, concurrency, payload size, memory budget, latency budget.
Identify hot paths and expensive dependencies: database, network, filesystem, CPU, serialization, rendering, queues.
Run signal scan:

python3 scripts/performance_signal_scan.py /path/to/repo-or-file

Trace loops, queries, API calls, and allocations through realistic worst-case inputs.
Recommend validation: targeted tests, profiling, explain plans, benchmarks, load tests, or production metrics.

Output Rules

Findings must include scale assumptions and the cost mechanism.
Distinguish measured issues from plausible risks.
Prefer concrete mitigations: add pagination, batch calls, add index, cache with invalidation, stream data, avoid repeated work.
If no findings are found, state residual risk and whether profiling or benchmarks were skipped.