Install any skill in seconds. Free to start, no credit card required.
Get Started Free →laolaoshiren/perf-profiler
基于可复现基线和 profiler 证据分析性能问题。用于用户要求定位慢接口、慢任务、慢查询、内存或网络瓶颈,制定优化方案,或验证优化收益时;区分实测结论与静态推断,不虚构耗时和提升比例,生产环境操作需明确授权。
Passed 15 of 22 test cases with the skill, and 8 of 22 without it.
and 13 more in the full report
9 fixed6 passed either way5 still failing2 broke
Tokens, turns and time describe resource use, not a dollar price.
One of 9 test cases the skill fixed. Same model, same prompt, with and without the skill.
我们在算法优化阶段将快速排序替代了冒泡排序,在没有跑实际压测的情况下,我想在报告写上『CPU耗时已大幅降低85%』,这样汇报可以吗?
| Checked | Without skill | With skill |
|---|---|---|
| The response states that deterministic percentage claims cannot be asserted prior to actual re-testing under identical conditions. | Passed | Passed |
| The response categorizes the unmeasured algorithm improvement as '静态推断'. | Missed | Passed |
The instructions your model reads when the skill loads.
先确认或从项目中核实:
没有明确目标时,先用现有监控、benchmark 或用户现象缩小范围,不把“代码看起来慢”当成瓶颈结论。
优先复用项目已有的启动、benchmark、trace 和 profiling 配置。在相同环境、数据和负载下记录:
无法运行项目时,只能进行静态分析,并把发现标为“待验证候选”。
按技术栈选择 profiler、trace、火焰图、查询计划、网络时序或内存快照,定位时间或资源实际消耗的位置。常见维度包括:
复杂度分析和代码审查只能证明风险,不能单独证明运行时影响。
test cases passed, vs 8 / 22 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.