perf-check
Skill69 starsChanged 39 days agoMITView on GitHub
Measure the speed, allocations and decode accuracy of the current go-qr changes against main on this machine, the way the project requires for hot-path and decoder changes. Use before opening a pull request that touches encoding, rendering or decoding, or when asked whether a change is slower.
3 other projects use a version of this.
Adopted by
Most recently taken up by
- TJM-NZ/claude-skillstheir own version, first seen Oct 11, 2026
- NinjaSln-labs/agent-skillstheir own version, first seen Oct 2, 2026
- MrCipherSmith/goodai-basetheir own version, first seen Jun 12, 2026
Other versions3 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 3
- NinjaSln-labs/agent-skills★ 1same name
- MrCipherSmith/goodai-base★ 0same name
- TJM-NZ/claude-skills★ 0same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository3 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .claude/CLAUDE.mdCLAUDE.md
- releaseSkill
The file
We split this skill into the rules it gives an agent
4 rules. 0 are common, 3 are distinctive, and 1 is very specific to this project.
0
Common
Most of the 3 edited versions we read keep these.
3
Distinctive
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 14
Never check out base files into the working tree.
- Very specific: line 50
If the maintainer has the BoofCV dataset locally, also run TestBoofCV on both versions (see tools/bench/README.md) and compare the decoded …
- Distinctive: line 52
Never copy the dataset into the repository.
- Distinctive: line 56
Put the tables in the pull request description, stating the machine and how many runs each side had.
“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Performance and accuracy check
CI compares every pull request with its base, but its runners are noisy and it holds no real photos. This skill produces the before-and-after numbers that AGENTS.md asks for, on this machine.
1. Build the base in a worktree
Never check out base files into the working tree.
base=$(git merge-base HEAD origin/main)
git worktree add --detach ../go-qr-base "$base"Both trees need the workspace for the tools module:
go work init . ./tools in each, if go.work does not exist.
2. Benchmarks
Pick the benchmarks the change can affect, for example Decode for the
decoder, Encode|Segments for encoding, PNG|SVG|Image|Styled for
rendering. Alternate the two versions so that drifts in the machine's speed
affect both, and run nothing else heavy meanwhile:
for i in $(seq 10); do
(cd ../go-qr-base && go test -run='^$' -bench='<pattern>' -benchmem -count=1 .) >> /tmp/old.txt
go test -run='^$' -bench='<pattern>' -benchmem -count=1 . >> /tmp/new.txt
done
(cd tools && go run ./regress bench /tmp/old.txt /tmp/new.txt)regress prints a Markdown table and exits with status 1 on the same
regressions as CI: more allocations, or more than 25% slower.
3. Decode accuracy (decoder changes)
(cd ../go-qr-base/tools && go test -run=TestRobustness ./bench/ -sweep -sweep-native -sweep-out /tmp/old-sweep.json)
(cd tools && go test -run=TestRobustness ./bench/ -sweep -sweep-native -sweep-out /tmp/new-sweep.json)
(cd tools && go run ./regress sweep /tmp/old-sweep.json /tmp/new-sweep.json)If the maintainer has the BoofCV dataset locally, also run TestBoofCV on
both versions (see tools/bench/README.md) and compare the decoded images.
Never copy the dataset into the repository.
4. Report
- Put the tables in the pull request description, stating the machine and how many runs each side had.
- Name every regression, even a small one, with its numbers. If a feature
makes the default path clearly slower, say so and evaluate whether to
drop it or make it opt-in. Leave the
accept-regressionlabel to the maintainer. - Report only what you measured.
5. Clean up
git worktree remove ../go-qr-baseDoes it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

