SlopBench: a frontend React/TypeScript slop benchmark (packages/benchmark)#757

Draft

NisargIO wants to merge 30 commits into

mainfrom

cursor/bc-c6c9d152-d3f7-49c9-baac-82f1c5d76a97-0bbb

Commits on Jun 9, 2026

feat(benchmark): scaffold SlopBench package + report contracts

cursoragent
and
NisargIO
committed
feat(benchmark): React Doctor scanner (offline, diff-scoped) + passing integration test

cursoragent
and
NisargIO
committed
feat(benchmark): deterministic local slop scorer + weights + golden tests

cursoragent
and
NisargIO
committed
feat(benchmark): TS-strictness + composition + deslop AST checks (gap-fillers) + rule-overlap doc

cursoragent
and
NisargIO
committed
feat(benchmark): slop-verify orchestrator + CLI + diff collection + e2e tests

cursoragent
and
NisargIO
committed
feat(benchmark): Harbor/Pier task harness — base image, shared grader, task _template

cursoragent
and
NisargIO
committed
feat(benchmark): task format-money-util (validated) + --no-dead-code in verifier + task validator

cursoragent
and
NisargIO
committed
feat(benchmark): task scaffolding + patch generators; add typed-storage-util task (validated)

cursoragent
and
NisargIO
committed
feat(benchmark): handle-slop task group-by-extend (validated)

cursoragent
and
NisargIO
committed
feat(benchmark): React/vitest task status-pill-variants (validated)

cursoragent
and
NisargIO
committed
feat(benchmark): notification-list task (validated; demonstrates react slop discrimination)

cursoragent
and
NisargIO
committed
feat(benchmark): format-duration-util task + validate-all script (all 6 tasks green)

cursoragent
and
NisargIO
committed
feat(benchmark): README + results aggregator + methodology docs; scope unit tests to tests/; format

cursoragent
and
NisargIO
committed
feat(benchmark): dashboard-loader async-waterfall task (Next server loader; validated)

cursoragent
and
NisargIO
committed
feat(benchmark): add slugify, parse-query, format-list-extend tasks (10 tasks; all validated)

cursoragent
and
NisargIO
committed
feat(benchmark): add truncate-middle + avatar-initials tasks (12 tasks; all validated)

cursoragent
and
NisargIO
committed
ci(benchmark): wire SlopBench task self-test into CI + add benchmark:validate script

cursoragent
and
NisargIO
committed
docs(benchmark): reference pnpm benchmark:validate in README

cursoragent
and
NisargIO
committed
feat(benchmark): icon-button-a11y task (covers accessibility dimension; 13 tasks; validated)

cursoragent
and
NisargIO
committed
feat(benchmark): chunk-util task (covers bundle dimension; 14 tasks; all validated)

cursoragent
and
NisargIO
committed
feat(benchmark): route-handler-json Next App Router task (query filtering; 15 tasks; validated)

cursoragent
and
NisargIO
committed
feat(benchmark): comment-thread-extend handle-slop React task (16 tasks; validated)

cursoragent
and
NisargIO
committed
feat(benchmark): add paginate-util + retry-async-util tasks (18 tasks; all validated, references pristine)

cursoragent
and
NisargIO
committed
feat(benchmark): add unique-by + title-case tasks (20 tasks; full agreed range; all references pristine)

cursoragent
and
NisargIO
committed
fix(benchmark): regenerate task patches after formatting so all 20 apply cleanly

cursoragent
and
NisargIO
committed
docs(benchmark): note to format before generating task patches

cursoragent
and
NisargIO
committed
docs(benchmark): clean up authoring note (pnpm format before gen-task-patches)

cursoragent
and
NisargIO
committed
test(benchmark): cover results aggregator (pass-rate, means, per-dimension, empty logs)

cursoragent
and
NisargIO
committed
fix(benchmark): raise vitest timeout for React Doctor-spawning integration tests (CI 5s default timed out)

cursoragent
and
NisargIO
committed
fix(docs): format docs/SLOPBENCH.md (CI format:check)

cursoragent
and
NisargIO
committed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

SlopBench: a frontend React/TypeScript slop benchmark (packages/benchmark)#757

SlopBench: a frontend React/TypeScript slop benchmark (packages/benchmark)#757
NisargIO wants to merge 30 commits into
mainfrom
cursor/bc-c6c9d152-d3f7-49c9-baac-82f1c5d76a97-0bbb

Commits on Jun 9, 2026

feat(benchmark): scaffold SlopBench package + report contracts

feat(benchmark): React Doctor scanner (offline, diff-scoped) + passing integration test

feat(benchmark): deterministic local slop scorer + weights + golden tests

feat(benchmark): TS-strictness + composition + deslop AST checks (gap-fillers) + rule-overlap doc

feat(benchmark): slop-verify orchestrator + CLI + diff collection + e2e tests

feat(benchmark): Harbor/Pier task harness — base image, shared grader, task _template

feat(benchmark): task format-money-util (validated) + --no-dead-code in verifier + task validator

feat(benchmark): task scaffolding + patch generators; add typed-storage-util task (validated)

feat(benchmark): handle-slop task group-by-extend (validated)

feat(benchmark): React/vitest task status-pill-variants (validated)

feat(benchmark): notification-list task (validated; demonstrates react slop discrimination)

feat(benchmark): format-duration-util task + validate-all script (all 6 tasks green)

feat(benchmark): README + results aggregator + methodology docs; scope unit tests to tests/; format

feat(benchmark): dashboard-loader async-waterfall task (Next server loader; validated)

feat(benchmark): add slugify, parse-query, format-list-extend tasks (10 tasks; all validated)

feat(benchmark): add truncate-middle + avatar-initials tasks (12 tasks; all validated)

ci(benchmark): wire SlopBench task self-test into CI + add benchmark:validate script

docs(benchmark): reference pnpm benchmark:validate in README

feat(benchmark): icon-button-a11y task (covers accessibility dimension; 13 tasks; validated)

feat(benchmark): chunk-util task (covers bundle dimension; 14 tasks; all validated)

feat(benchmark): route-handler-json Next App Router task (query filtering; 15 tasks; validated)

feat(benchmark): comment-thread-extend handle-slop React task (16 tasks; validated)

feat(benchmark): add paginate-util + retry-async-util tasks (18 tasks; all validated, references pristine)

feat(benchmark): add unique-by + title-case tasks (20 tasks; full agreed range; all references pristine)

fix(benchmark): regenerate task patches after formatting so all 20 apply cleanly

docs(benchmark): note to format before generating task patches

docs(benchmark): clean up authoring note (pnpm format before gen-task-patches)

test(benchmark): cover results aggregator (pass-rate, means, per-dimension, empty logs)

fix(benchmark): raise vitest timeout for React Doctor-spawning integration tests (CI 5s default timed out)

fix(docs): format docs/SLOPBENCH.md (CI format:check)