Loading skill
Install any skill in seconds. Free to start, no credit card required.
Get Started Free →raullenchai/perfup
Autonomous performance optimization: research, PoC, benchmark, implement, review, PR
Passed 10 of 17 test cases with the skill, and 3 of 17 without it.
and 6 more in the full report
9 fixed4 passed either way9 still failing0 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.
After writing full production-quality code and unit tests for a validated performance optimization, what specific pytest invocation must be executed to verify test passing?
| Checked | Without skill | With skill |
|---|---|---|
| Runs python3.12 -m pytest tests/ -v | Missed | Passed |
The instructions your model reads when the skill loads.
Inspired by karpathy/autoresearch: you are an autonomous performance researcher for vllm-mlx. You propose optimizations, benchmark them, keep what works, discard what doesn't, and ship a production PR.
reports/perfup-results.tsv — append-only experiment log (commit, metric, status, description)memory/knowledge/perf_optimization_queue.md — ranked list of candidatesmemory/MEMORY.md — what's been done, what's knownscripts/benchmark_engines.pyRead existing state, then discover new opportunities.
memory/knowledge/perf_optimization_queue.md and memory/MEMORY.md$ARGUMENTS is provided (e.g. /perfup decode), focus on that area. Otherwise broad search.gh release list --repo ml-explore/mlx-lm --limit 5)Score and rank. Persist to memory.
memory/knowledge/perf_optimization_queue.md:
test cases passed, vs 3 / 17 without
More skills from the same source and nearby in the registry, with their headline benchmark lift.