Find where the time actually goes before optimising: sampling versus instrumentation, reading a flame graph, and separating a slow algorithm from a slow I/O wait.
NewMaintained
Automatic
◆Coding pack
Last verified 8/3/2026
What it changes
Passed 21 of 22 test cases with the skill, and 21 of 22 without it.
Without skill95%
With skill95%
Checks it passed that the model alone missed
Explains how to distinguish CPU execution from I/O wait by comparing process CPU time against wall-clock duration.
Explains that sampling periodically captures stack snapshots, whereas instrumentation hooks function entry and exit.
1 fixed20 passed either way0 still failing1 broke
Pass-rate lift
0 pp
Tokens
+3%
Agent turns
0%
Response time
-1%
Model gemini-3.6-flashMeasured 8/3/2026v121 of 22 cases passed with skill
Tokens, turns and time describe resource use, not a dollar price.
See it in action
A test case the skill fixed. Same model, same prompt, with and without the skill.
Prompt
My backend service is experiencing high latency, and I'm not sure if the bottleneck is caused by heavy computations or database queries. Can you walk me through how to profile the application, interpret flame graphs, and differentiate CPU-bound execution from I/O wait times before I start refactoring?
What was checked, without and with the skill
Checked
Without skill
With skill
Explains that sampling periodically captures stack snapshots, whereas instrumentation hooks function entry and exit.
Missed
Passed
Notes that the x-axis of a flame graph represents sample frequency rather than a chronological timeline.
Passed
Passed
Explains how to distinguish CPU execution from I/O wait by comparing process CPU time against wall-clock duration.
Missed
Passed
case-01 · gemini-3.6-flash · 8/3/2026 · answers are capped and redacted
What's inside
The instructions your model reads when the skill loads.
SKILL.md5 linessingle file
perf-profiling
Find where the time actually goes before optimising: sampling versus
instrumentation, reading a flame graph, and separating a slow algorithm from a
slow I/O wait.