191 lines
53 KiB
JSON
191 lines
53 KiB
JSON
{"prompt": "design the offline sync layer for Kestrel, our SwiftUI notes app. i need a conflict resolution story for concurrent edits on iPad + Mac, plus a migration path off the current UserDefaults blob. write it up as an RFC with the tradeoffs of CRDT vs last-write-wins spelled out", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.85, "slice": "core", "lang": "en"}
|
|
{"prompt": "add a GET /v1/exports endpoint in Go that streams NDJSON out of the reports table, cursor paginated on (created_at, id)", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
|
{"prompt": "make the toolbar collapse on scroll", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "core", "lang": "en"}
|
|
{"prompt": "bump golang.org/x/net to 0.38", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "core", "lang": "en"}
|
|
{"prompt": "SessionStore.swift is up to 900 lines. pull the keychain access and the token refresh timer out into their own types, same behavior after", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.55, "slice": "core", "lang": "en"}
|
|
{"prompt": "crashes on launch for a handful of users on 18.4, never in the simulator. crashlytics gives me this and i'm lost:\n\nFatal Exception: NSInternalInconsistencyException\nModifying state during view update, this will cause undefined behavior.\n0 SwiftUI 0x1a2 closure #1 in AttributeGraph.withMutation + 412\n1 SwiftUI 0x1a2 StateBox.setValue(_:transaction:) + 88\n2 Kestrel 0x104 LibraryViewModel.reload() + 240 (LibraryViewModel.swift:118)\n3 Kestrel 0x104 closure #2 in LibraryView.body.getter + 96 (LibraryView.swift:64)\n4 SwiftUI 0x1a2 ViewGraph.updateOutputs + 1204\n5 UIKitCore 0x18f _UIUpdateSequenceRun + 84\n6 CoreFoundation 0x180 __CFRUNLOOP_IS_CALLING_OUT_TO_AN_OBSERVER_CALLBACK_FUNCTION__ + 36\n\nLibraryViewModel.reload() is called from .task and from a NotificationCenter observer for .didFinishImport. no idea why only device builds", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.75, "slice": "pasted-context", "lang": "en"}
|
|
{"prompt": "walk me through how our terraform module wires ALB target groups to the ECS service — mainly i want to know who actually owns the health check config, the tg or the task def", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.45, "slice": "core", "lang": "en"}
|
|
{"prompt": "draft release notes for 3.4 off the merged PRs since the v3.3 tag, user facing tone, group by area", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.35, "slice": "core", "lang": "en"}
|
|
{"prompt": "how would we get off single-node postgres onto citus for the events table? i don't want code yet, just the phases, what has to be dual-written, and which of our queries break when there's no global sequence", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "our openapi.yaml has bare field names and nothing else for /webhooks/*. fill in descriptions and realistic examples for every field and document the retry/backoff semantics we already implement. don't touch the handlers", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "generate openapi.yaml from the existing gin route handlers and struct tags, and wire swag into the make target so it regenerates on build", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.55, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "empty state for the workspace picker — illustration, one line of copy, a Create workspace button. match the spacing of the other empties", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.35, "slice": "core", "lang": "en"}
|
|
{"prompt": "rename the local var `resp` to `httpResp` in fetchInvoices, it shadows the outer one and it's confusing", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "we're calling it Organization everywhere in the API but Tenant in the db layer and Account in a few older services. standardize on Organization across the whole repo including column aliases and proto field comments, no functional change", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "the nightly ingest job silently processes 0 rows about one run in five, no error logged, no alert. figure out what's happening", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
|
{"prompt": "why does the shipment reducer dispatch twice for a single scan event? not saying it's wrong, i just want to understand the flow before i touch it", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.45, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "plan and then build the rate limiter for our public API — per-key token bucket, redis backed, needs to survive a redis blip without locking everyone out. give me the design first then the go code", "purpose": "planning", "secondary": "backendImpl", "mixed": true, "difficulty": 0.75, "slice": "mixed", "lang": "en"}
|
|
{"prompt": "commit message for these staged changes pls", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "core", "lang": "en"}
|
|
{"prompt": "milestones for the Q3 auth rewrite", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
|
{"prompt": "need a background worker in python that watches the s3 bucket prefix raw/telemetry/, batches parquet files into 15 minute windows, and upserts into clickhouse. idempotent on rerun, and it has to handle files landing out of order up to 2h late. use the existing boto session helper in common/aws.py", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
|
{"prompt": "sketch the interaction model for the new inline comment threads (hover affordance, focus states, keyboard nav) and then implement the React side. no backend, the endpoints exist", "purpose": "planning", "secondary": "frontendImpl", "mixed": true, "difficulty": 0.7, "slice": "mixed", "lang": "en"}
|
|
{"prompt": "feature flag `new_billing_summary` is still gated to staff only, open it to everyone", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.15, "slice": "core", "lang": "en"}
|
|
{"prompt": "dedupe the three copies of retryWithBackoff", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.35, "slice": "core", "lang": "en"}
|
|
{"prompt": "the /search endpoint got 4x slower after we shipped last thursday and nothing in that diff touches search. p99 went 180ms -> 750ms. where do i even start", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.75, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "make it pop", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "vague-eval", "lang": "en"}
|
|
{"prompt": "is our jwt refresh flow safe against replay if someone grabs a refresh token off a stolen device? read through auth/refresh.go and tell me what you think, don't change anything yet", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.65, "slice": "core", "lang": "en"}
|
|
{"prompt": "compare our current celery setup against just using postgres SKIP LOCKED for the job queue. we do maybe 200 jobs/min, mostly short", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
|
{"prompt": "explain what pkg/ledger/reconcile.go actually does, line by line if you have to. inherited it and nobody left knows", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.55, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "summarize the reconcile flow for the onboarding doc — same file, but i want three paragraphs a new hire can read, not a code walkthrough", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.45, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "roadmap for deprecating the v1 REST API. we have 40ish external integrators, 9 months of runway, and i need something i can show the CTO on monday", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
|
{"prompt": "how should we shard the notification fanout? design only", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
|
{"prompt": "we want an audit log for every admin mutation. figure out the approach — trigger based, app level, or CDC off the wal — and write up which one you'd pick and why before anyone writes code", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.75, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "implement soft delete on the projects table: deleted_at column, partial unique index on slug where deleted_at is null, and update every query in internal/projects to filter it. also a nightly purge after 30 days", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
|
{"prompt": "write a small CLI in rust, `envdiff`, that takes two .env files and prints keys only present in one, keys whose values differ, and exits 1 if there's any drift", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.45, "slice": "core", "lang": "en"}
|
|
{"prompt": "webhook signature verification for the stripe events handler, constant time compare, tolerate 5 min clock skew", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.45, "slice": "core", "lang": "en"}
|
|
{"prompt": "dark mode for the settings sheet is unreadable — the secondary labels are basically the same value as the background. go through Settings/ and fix the color roles properly instead of hardcoding hex", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.45, "slice": "core", "lang": "en"}
|
|
{"prompt": "add a skeleton loader to the analytics cards", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.35, "slice": "core", "lang": "en"}
|
|
{"prompt": "the primary button has 14px padding on web and 12 on the marketing site. make them both 14", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "typo in the onboarding modal, 'Contine' -> 'Continue'", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.05, "slice": "core", "lang": "en"}
|
|
{"prompt": "set the readiness probe initialDelaySeconds to 20 in the api deployment, it's flapping on cold start", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "core", "lang": "en"}
|
|
{"prompt": "break InvoiceService apart. right now it does pdf rendering, tax calc, and the stripe calls in one class and i can't test any of it in isolation. keep the public methods identical so callers don't change", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.65, "slice": "core", "lang": "en"}
|
|
{"prompt": "collapse our four nearly identical Dockerfiles into one multi-stage with build args", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
|
{"prompt": "convert the callback-style code in lib/uploader.js to async/await, don't change any behavior including the weird double-callback guard", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
|
{"prompt": "flaky test, fails maybe 1 in 12 locally and more in CI:\n\n=================================== FAILURES ===================================\n_______________________ test_expiry_sweeper_removes_stale ______________________\n\n async def test_expiry_sweeper_removes_stale(session, freezer):\n await create_lease(session, ttl=30)\n freezer.move_to('+45s')\n await sweeper.run_once(session)\n> assert await count_leases(session) == 0\nE assert 1 == 0\nE + where 1 = <coroutine count_leases>\n\ntests/test_sweeper.py:88: AssertionError\n------------------------------ Captured log call -------------------------------\nDEBUG app.sweeper:sweeper.py:41 sweep start cutoff=2026-03-02T10:00:45+00:00\nDEBUG app.sweeper:sweeper.py:57 candidate leases=1\nDEBUG app.sweeper:sweeper.py:63 skipped lease id=8812 reason=locked\n=========================== short test summary info ============================\nFAILED tests/test_sweeper.py::test_expiry_sweeper_removes_stale - assert 1 == 0\n1 failed, 213 passed, 4 skipped in 61.42s\n\nthe 'locked' skip only shows up on the failing runs", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "pasted-context", "lang": "en"}
|
|
{"prompt": "memory climbs about 40MB/hour in the websocket gateway until the pod gets oomkilled overnight. connections are stable around 3k. help me find the leak", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
|
{"prompt": "gradle build hangs at :app:kaptDebugKotlin, no output, have to ctrl-c after 10 min. worked yesterday", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.65, "slice": "core", "lang": "en"}
|
|
{"prompt": "PR description for the branch feat/multi-region-reads, include the migration steps ops needs to run and the rollback", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
|
{"prompt": "write doc comments for every exported symbol in pkg/schedule, godoc style, mention the timezone caveat on NextRun", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "our README still says node 16 and describes the old yarn workspaces setup. rewrite the getting started section for the current pnpm + turbo layout", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.35, "slice": "core", "lang": "en"}
|
|
{"prompt": "escribe la documentación del endpoint POST /pedidos en el openapi, incluyendo los códigos de error y un ejemplo de payload real", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.35, "slice": "boundary", "lang": "es"}
|
|
{"prompt": "changelog entry for the 2.1.7 patch, we fixed the timezone bug in recurring reminders and dropped support for android 8", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "core", "lang": "en"}
|
|
{"prompt": "translate the error strings in Localizable.strings to de and fr, keep the format specifiers in the same order", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
|
{"prompt": "架构一下我们的实时协作编辑，需要支持离线编辑和多端同步，先给方案不要写代码", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "zh"}
|
|
{"prompt": "we keep firefighting the same three classes of prod incident. help me put together a plan for the next quarter that actually reduces them — my notes: no request tracing across services, alerts fire on symptoms not causes, and nobody owns the canary step", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
|
{"prompt": "how do i structure a plugin system for the editor so third parties can add commands without us shipping their code in the bundle? want the design tradeoffs, esp sandboxing", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.85, "slice": "core", "lang": "en"}
|
|
{"prompt": "add cursor based pagination to the ListDevices grpc method, keep the old offset params working for one release", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.55, "slice": "core", "lang": "en"}
|
|
{"prompt": "parse the vendor's fixed width inventory feed. spec attached below in the ticket text, 320 byte records, EBCDIC in a couple of the numeric fields apparently. go, streaming, don't load the whole file", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
|
{"prompt": "wire up SSO with our okta tenant — SAML, JIT provisioning into the users table, and map the groups claim onto our existing roles", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.75, "slice": "core", "lang": "en"}
|
|
{"prompt": "drag to reorder on the playlist rows, with haptics", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
|
{"prompt": "build the compare view: two device timelines side by side, shared x axis, scrub together, and it needs to stay smooth at 5k points each. we're on SwiftUI Charts", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.75, "slice": "core", "lang": "en"}
|
|
{"prompt": "nudge the badge 2px left on the tab bar", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "prod is on log level debug, drop it to info", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "core", "lang": "en"}
|
|
{"prompt": "the retry count is 3 but the ticket says it should be 5 for the payout webhook. one line, i know where it is, worker/payouts.ts around line 70", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "add .DS_Store and .idea/ to gitignore and untrack whatever's already in", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.15, "slice": "core", "lang": "en"}
|
|
{"prompt": "extract the validation out of every handler into middleware. it's copy pasted in 22 places and i want them all going through one path with the same error shape", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.65, "slice": "core", "lang": "en"}
|
|
{"prompt": "kill the God object. AppState in our react-native app holds auth, cart, feature flags, and the socket. split by domain, keep the selector API stable for now", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
|
{"prompt": "move the shared types into a packages/contracts workspace so the api and the web app stop importing across each other's src/", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
|
{"prompt": "getting this on every deploy to the eu cluster only:\n\nEvents:\n Type Reason Age From Message\n ---- ------ ---- ---- -------\n Normal Scheduled 4m12s default-scheduler Successfully assigned prod/api-7d9c8-hlq2n to ip-10-3-42-118\n Normal Pulled 4m10s kubelet Container image \"registry.internal/api:9f2c1a\" already present on machine\n Normal Created 4m10s kubelet Created container api\n Normal Started 4m9s kubelet Started container api\n Warning Unhealthy 3m40s (x6 over 4m) kubelet Readiness probe failed: Get \"http://10.3.42.7:8080/readyz\": context deadline exceeded\n Warning BackOff 2m1s (x4 over 3m) kubelet Back-off restarting failed container api\n\nand in the container logs:\nlevel=info msg=\"connecting to pgbouncer\" host=pgb.eu-central.internal\nlevel=warn msg=\"tls handshake retry\" attempt=1\nlevel=warn msg=\"tls handshake retry\" attempt=2\nlevel=fatal msg=\"x509: certificate is valid for pgb.us-east.internal, not pgb.eu-central.internal\"\n\nus-east is totally fine which is what confuses me", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "pasted-context", "lang": "en"}
|
|
{"prompt": "someone dropped this in #eng-help and left for the day:\n\n> hey so the CSV export for org 44121 has been producing files where about 3% of rows have the amount column shifted one to the right. only that org. started sometime after the 12th. i checked and their data doesn't look special, no weird currency. the file opens fine in excel it's just wrong. customer is escalating\n\ncan you dig into it? export code is in services/reporting/csv.py", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.75, "slice": "pasted-context", "lang": "en"}
|
|
{"prompt": "audio glitches for ~200ms every time we switch bluetooth routes on the embedded player. only on the nRF52 build, not the sim", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
|
{"prompt": "is it a problem that we're doing the password hash on the request goroutine with argon2 at 64MB? worried about a thundering herd at login but i haven't measured anything", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.55, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "review the diff on my branch before i open the PR, specifically the transaction boundaries in the transfer path", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
|
{"prompt": "what's the difference between our useSyncedQuery hook and just using react-query's useQuery with our fetcher? feels like we reimplemented it", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.45, "slice": "core", "lang": "en"}
|
|
{"prompt": "document the event payload schemas for the pubsub topics in docs/events.md AND add the missing JSON schema files under schemas/ so we can validate in CI", "purpose": "writing", "secondary": "backendImpl", "mixed": true, "difficulty": 0.55, "slice": "mixed", "lang": "en"}
|
|
{"prompt": "continue", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "vague-eval", "lang": "en"}
|
|
{"prompt": "figure out why the checkout total is off by a cent for some carts and then write up the postmortem, we told the customer we'd have both by friday", "purpose": "debugging", "secondary": "writing", "mixed": true, "difficulty": 0.7, "slice": "mixed", "lang": "en"}
|
|
{"prompt": "picking the payment-flow bug back up from yesterday — the 3DS redirect lands on a blank page maybe a third of the time in safari. no console error, network tab shows the POST to /confirm returning 302 and then nothing", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.75, "slice": "core", "lang": "en"}
|
|
{"prompt": "come up with the migration plan for moving our 4TB mysql instance to aurora with under 15 min of write downtime, and list the go/no-go checks", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.9, "slice": "core", "lang": "en"}
|
|
{"prompt": "break the vector search feature into two week chunks for a team of three", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
|
{"prompt": "before we build anything, i want an ADR on whether the mobile clients talk to a BFF or straight to the services. we have ios, android, and a watch app coming", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.75, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "implement optimistic locking on the inventory adjust path with a version column, return 409 with the current version on conflict", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.55, "slice": "core", "lang": "en"}
|
|
{"prompt": "hook the feature flag SDK into our fastapi app so flags resolve once per request and get attached to the request state. also expose them on /debug/flags behind the admin guard", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
|
{"prompt": "write the migration + backfill for splitting `full_name` into first/last on the contacts table. 8M rows, can't lock it", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.65, "slice": "core", "lang": "en"}
|
|
{"prompt": "the achievement popup needs to feel good — slide up, slight overshoot, confetti burst, auto dismiss after 3s but pause if the player is holding a button. unity, we're using the UI toolkit", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
|
{"prompt": "responsive pass on the pricing page, it's broken between 768 and 900", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
|
{"prompt": "add a compact density option to the data grid — user setting, persists, rows go 44 -> 32 and the font drops a step", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
|
{"prompt": "swap the deprecated `substr` calls for slice in utils/text.ts, six spots", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.15, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "cors is missing the staging origin, add https://staging.acme.dev to the allowlist", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.15, "slice": "core", "lang": "en"}
|
|
{"prompt": "off by one in the pagination footer, it says 'showing 1-20 of 20' when there are 21. i know it's the `hasMore ? end : total` bit in Pager.tsx, just fix it", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "consolidate the two http clients. one's axios with our interceptors, one's raw fetch with a hand rolled retry. pick fetch, keep every call site's behavior identical including the odd 204 handling", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
|
{"prompt": "untangle the import cycle between internal/billing and internal/accounts", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
|
{"prompt": "simplify this state machine, there are 11 states and i'm pretty sure 4 are unreachable. behavior must not change for the reachable ones", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.65, "slice": "core", "lang": "en"}
|
|
{"prompt": "rust won't let me do the thing and i've been at it an hour:\n\nerror[E0502]: cannot borrow `self.buffer` as mutable because it is also borrowed as immutable\n --> src/decoder.rs:214:13\n |\n209 | let frame = self.peek_frame()?;\n | ---- immutable borrow occurs here\n...\n214 | self.buffer.advance(frame.len);\n | ^^^^^^^^^^^^^^^^^^^ mutable borrow occurs here\n...\n219 | Ok(frame)\n | ----- immutable borrow later used here\n\nerror[E0499]: cannot borrow `*self` as mutable more than once at a time\n --> src/decoder.rs:231:9\n |\n230 | let hdr = self.header_mut();\n | ---- first mutable borrow occurs here\n231 | self.reset_crc();\n | ^^^^ second mutable borrow occurs here\n\nerror: could not compile `wire-codec` (lib) due to 2 previous errors\n\npeek_frame returns a &Frame into the buffer. what's the right shape here", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "pasted-context", "lang": "en"}
|
|
{"prompt": "intermittent 502s from the api gateway, maybe 1 in 400 requests, no pattern i can see by endpoint or region. logs on our side show a clean 200", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
|
{"prompt": "warum liefert unser cache manchmal veraltete daten obwohl die TTL 30s ist? ich verstehe es nicht, redis sieht sauber aus", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "boundary", "lang": "de"}
|
|
{"prompt": "read through the terraform in infra/prod and tell me what would break if we lost the whole eu-west-1 region tomorrow", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.75, "slice": "core", "lang": "en"}
|
|
{"prompt": "does this query plan look sane to you? it's the dashboard aggregate, runs 400ms on 2M rows and i can't tell if that's good or bad for what it's doing", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "explain the difference in guarantees between our at-least-once sqs consumer and the kafka one, i keep confusing myself in design reviews", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
|
{"prompt": "internal blog post about the latency work we did last quarter. audience is other engineers here, ~800 words, i can give you the numbers", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.45, "slice": "core", "lang": "en"}
|
|
{"prompt": "runbook for the on-call rotation covering the four alerts that actually page us", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
|
{"prompt": "the CONTRIBUTING.md is 3 lines. write a real one — branch naming, how to run the test suite, what we expect in a PR, and the codegen step people always forget", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
|
{"prompt": "do the thing we discussed", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "vague-eval", "lang": "en"}
|
|
{"prompt": "spec out the multiplayer lobby then get the netcode skeleton in — matchmaking rules first as a doc, then the actual go server with the room state machine", "purpose": "planning", "secondary": "backendImpl", "mixed": true, "difficulty": 0.85, "slice": "mixed", "lang": "en"}
|
|
{"prompt": "evaluate whether we should adopt bazel. monorepo, ~60 services, mixed go/ts/python, current builds are makefiles and 14 min of CI. i want a recommendation with a phased rollout, not a yes/no", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.85, "slice": "core", "lang": "en"}
|
|
{"prompt": "what's a reasonable retention + partitioning strategy for the raw telemetry table? we ingest ~90GB/day and only ever query the last 14 days interactively", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
|
{"prompt": "write the presigned upload flow: POST /uploads returns a presigned put + an id, then a lambda on s3:ObjectCreated validates the mime type and marks the row ready. reject anything over 25MB, and clean up rows that never get an object after 24h", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
|
{"prompt": "give the CLI a `--json` output mode across all subcommands, stable field names, and make sure nothing else writes to stdout when it's set", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.55, "slice": "core", "lang": "en"}
|
|
{"prompt": "batch the embedding calls in the indexer, 64 at a time, with a semaphore so we don't blow the rate limit. retry the whole batch on 429 with jitter", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
|
{"prompt": "sticky column headers in the transactions table, and the horizontal scroll shadow like linear has", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
|
{"prompt": "the map pins overlap badly when zoomed out. cluster them", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.55, "slice": "core", "lang": "en"}
|
|
{"prompt": "make the loading spinner smaller on mobile", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "increase the lambda timeout to 60s and memory to 1024", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "core", "lang": "en"}
|
|
{"prompt": "we're pinning pydantic 1.10 in requirements.txt but the code is v2 style. change the pin", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "core", "lang": "en"}
|
|
{"prompt": "the healthcheck path in nginx.conf points at /health but the app serves /healthz. one char fix basically", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "core", "lang": "en"}
|
|
{"prompt": "pull the pure functions out of the OrderViewModel so i can unit test them without a view. no logic changes", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
|
{"prompt": "we have `getUserById`, `fetchUser`, `loadUserRecord` and `userLookup` all doing the same query with slightly different caching. unify them", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "replace the stringly typed status field with a proper enum everywhere, including the db check constraint and the ts union", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
|
{"prompt": "espresso test keeps failing on CI but passes on my machine every time. IdlingResource is registered. anyone would think it's timing but i've bumped the timeouts to 30s and it still fails at the same assertion", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
|
{"prompt": "npm ci exploded on the release build:\n\nnpm error code ERESOLVE\nnpm error ERESOLVE could not resolve\nnpm error\nnpm error While resolving: @acme/[email protected]\nnpm error Found: [email protected]\nnpm error node_modules/react\nnpm error react@\"^19.1.0\" from the root project\nnpm error\nnpm error Could not resolve dependency:\nnpm error peer react@\"^18.2.0\" from @acme/[email protected]\nnpm error node_modules/@acme/charts\nnpm error @acme/charts@\"^2.9.0\" from @acme/[email protected]\nnpm error\nnpm error Conflicting peer dependency: [email protected]\nnpm error node_modules/react\nnpm error peer react@\"^18.2.0\" from @acme/[email protected]\nnpm error\nnpm error Fix the upstream dependency conflict, or retry\nnpm error this command with --force or --legacy-peer-deps\nnpm error A complete log of this run can be found in: /home/runner/.npm/_logs/2026-04-11T02_14_08_112Z-debug.log\n\nlocal install was fine last week and nothing in our package.json changed", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "pasted-context", "lang": "en"}
|
|
{"prompt": "why do we get duplicate rows in the nightly rollup only on the first of the month", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "security look at the file upload path — internal/media/upload.go plus the nginx config. what can a hostile client do", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
|
{"prompt": "is there any reason this uses a mutex instead of a channel here? genuinely asking, the code works fine", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "read the PR from the contractor (#4471) and give me an honest assessment. it's 2800 lines and touches the scheduler", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
|
{"prompt": "api reference page for the new /v2/search endpoint — params, sort syntax, the facet response shape, rate limits, and a couple of curl examples", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.45, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "summarize this week's incidents into something i can paste in the weekly update. i'll give you the three tickets", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "core", "lang": "en"}
|
|
{"prompt": "migration guide for integrators moving off the v1 webhooks — field mapping table, the signature change, and a timeline", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
|
{"prompt": "clean up and also make it faster", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "vague-eval", "lang": "en"}
|
|
{"prompt": "rename `Widget` to `Panel` throughout the design system package and then update the docs site examples to match", "purpose": "refactor", "secondary": "writing", "mixed": true, "difficulty": 0.55, "slice": "mixed", "lang": "en"}
|
|
{"prompt": "audit our dependency tree for anything unmaintained or with known CVEs and give me the list ranked by how exposed we actually are, then patch the trivial ones", "purpose": "review", "secondary": "quickFix", "mixed": true, "difficulty": 0.65, "slice": "mixed", "lang": "en"}
|
|
{"prompt": "propose the schema for multi-tenant row level security in postgres. we currently do it in the app layer with a where clause and i don't trust it", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "we need a story for feature flags at build time on embedded. 200KB of flash left, no network. options?", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
|
{"prompt": "outline the steps to get our monorepo CI under 5 minutes", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
|
{"prompt": "implement the delta sync endpoint: client sends a cursor, server returns changed + deleted ids since then, tombstones kept 30 days. also needs to work when the client's cursor is older than the tombstone window", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
|
{"prompt": "build a tiny http server in zig that serves a directory with range requests and etags. no deps", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
|
{"prompt": "put a dead letter queue behind the notification consumer, 5 attempts then park, plus a small admin endpoint to replay a parked message by id", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.55, "slice": "core", "lang": "en"}
|
|
{"prompt": "keyboard shortcuts overlay, cmd+/ to open, grouped by section, and it should reflect the actual registered commands not a hardcoded list", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.55, "slice": "core", "lang": "en"}
|
|
{"prompt": "the onboarding carousel doesn't respect reduce motion. also the page dots are too small to tap", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
|
{"prompt": "ajoute un mode sombre au tableau de bord, en respectant les tokens de couleur existants", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.45, "slice": "core", "lang": "fr"}
|
|
{"prompt": "regenerate the protobuf stubs, someone committed stale ones", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "core", "lang": "en"}
|
|
{"prompt": "the cron is set to 0 3 * * * but ops wants it at 3am pacific not utc, and the container has no TZ set. fix both", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "core", "lang": "en"}
|
|
{"prompt": "drop the console.log i left in checkout.tsx", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.05, "slice": "core", "lang": "en"}
|
|
{"prompt": "decouple the pdf renderer from the web request lifecycle — it's currently reaching into flask's g for the tenant and i want it to take an explicit context object", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
|
{"prompt": "the test helpers have drifted into three overlapping factory modules. merge them, keep every existing test passing untouched", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
|
{"prompt": "flatten the four levels of inheritance under BaseRepository into composition", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
|
{"prompt": "TODO list from the tech debt session, can you work through it:\n\n- [ ] `PaymentGateway` and `PaymentsGateway` both exist, one is a thin wrapper around the other. collapse\n- [ ] the `legacy_` prefix on 11 methods in BillingService means nothing anymore, drop the prefix\n- [ ] move `money.rb` out of app/lib into a proper Money value object under app/values\n- [ ] every controller has its own `current_org` memo, hoist to ApplicationController\n- [ ] `Invoice#total`, `Invoice#total_cents`, `Invoice#grand_total` — pick one, keep cents as the source of truth\n\nnone of this should change what the app does, our specs should stay green as-is", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "pasted-context", "lang": "en"}
|
|
{"prompt": "profiler says 60% of wall time is in json marshal on the hot path but the payloads are tiny. something's off, help me understand what i'm looking at", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "user reports the timer keeps running after they background the app on android 14 but the notification disappears. can't reproduce on my pixel", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
|
{"prompt": "sqlite db is getting corrupted on the field devices, maybe 1 in 50 units after a power cut. WAL is on. what am i doing wrong", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.85, "slice": "core", "lang": "en"}
|
|
{"prompt": "here's the diff, tell me if the locking is right:\n\n@@ -142,17 +142,29 @@ func (s *Scheduler) Claim(ctx context.Context, workerID string) (*Job, error) {\n-\ts.mu.Lock()\n-\tdefer s.mu.Unlock()\n-\n-\tjob := s.queue.Pop()\n-\tif job == nil {\n-\t\treturn nil, ErrNoJobs\n-\t}\n+\ts.mu.RLock()\n+\tjob := s.queue.Peek()\n+\ts.mu.RUnlock()\n+\n+\tif job == nil {\n+\t\treturn nil, ErrNoJobs\n+\t}\n+\n+\ts.mu.Lock()\n+\tif s.queue.Peek() != job {\n+\t\ts.mu.Unlock()\n+\t\treturn s.Claim(ctx, workerID)\n+\t}\n+\ts.queue.Pop()\n+\ts.mu.Unlock()\n \n \tjob.WorkerID = workerID\n \tjob.ClaimedAt = time.Now()\n-\treturn job, s.store.Update(ctx, job)\n+\tif err := s.store.Update(ctx, job); err != nil {\n+\t\ts.mu.Lock()\n+\t\ts.queue.Push(job)\n+\t\ts.mu.Unlock()\n+\t\treturn nil, err\n+\t}\n+\treturn job, nil\n }\n\nnot asking you to change it, i want to know if the retry path can spin or lose a job", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.75, "slice": "pasted-context", "lang": "en"}
|
|
{"prompt": "does our GDPR delete actually delete everything? trace a user id through every store including the analytics warehouse and the s3 exports", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.75, "slice": "core", "lang": "en"}
|
|
{"prompt": "what does the `-tags fastpath` build actually change in this repo", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
|
{"prompt": "deprecation notice for the SDK's `Client.legacyUpload` — what to use instead, when it goes away, and a before/after snippet", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "リリースノートを日本語で書いて、v5.2の変更点は認証周りの修正とパフォーマンス改善", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "core", "lang": "ja"}
|
|
{"prompt": "our error codes are documented nowhere. go through the codebase, collect every code we return, and produce a reference table with meaning + suggested client action", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.55, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "design the permissions model for shared folders — inheritance, explicit deny, link sharing — and then implement the check function once we're happy with the model", "purpose": "planning", "secondary": "backendImpl", "mixed": true, "difficulty": 0.85, "slice": "mixed", "lang": "en"}
|
|
{"prompt": "what's the plan for supporting airplane mode in the field app", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
|
{"prompt": "i want to introduce typed errors across our python services. write up the convention and the rollout order so 12 teams can adopt it without a big bang", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.75, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "add a /metrics endpoint with prometheus counters for request count, latency histogram by route, and queue depth", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
|
{"prompt": "implementa el login con refresh tokens rotativos en el servicio de auth, con revocación por dispositivo", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.65, "slice": "core", "lang": "es"}
|
|
{"prompt": "graphql resolver for `organization.usage(period:)` that hits the materialized view, with dataloader batching so we don't n+1 across the org list", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
|
{"prompt": "virtualize the log viewer, it chokes past 10k lines. keep the find-in-page highlight working", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
|
{"prompt": "give the empty chart state something better than 'No data'", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "core", "lang": "en"}
|
|
{"prompt": "sidebar should remember collapsed state per workspace", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
|
{"prompt": "set `strictNullChecks` true in tsconfig — just the flag, i'll deal with the fallout separately", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.15, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "the s3 bucket policy is missing the deny-insecure-transport statement, add it", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.25, "slice": "core", "lang": "en"}
|
|
{"prompt": "our slack alert says 'CPU is high' with no service name. include $SERVICE and the runbook link in the template", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "core", "lang": "en"}
|
|
{"prompt": "hoist the duplicated auth setup out of all 40 test files into a shared fixture", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
|
{"prompt": "the Kotlin data layer has repository interfaces that are only implemented once and add nothing. inline them and delete the abstraction", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.55, "slice": "core", "lang": "en"}
|
|
{"prompt": "port utils/date.js to typescript, same exports, no behavior change", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "ci went red on main and the output is useless to me:\n\n$ go test ./... -race -count=1\n==================\nWARNING: DATA RACE\nWrite at 0x00c0004b2118 by goroutine 87:\n github.com/acme/relay/internal/hub.(*Hub).register()\n /src/internal/hub/hub.go:132 +0x64\n github.com/acme/relay/internal/hub.(*Hub).Run.func2()\n /src/internal/hub/hub.go:88 +0x3c\n\nPrevious read at 0x00c0004b2118 by goroutine 12:\n github.com/acme/relay/internal/hub.(*Hub).Broadcast()\n /src/internal/hub/hub.go:201 +0x118\n github.com/acme/relay/internal/api.(*Server).handlePublish()\n /src/internal/api/publish.go:57 +0x244\n\nGoroutine 87 (running) created at:\n github.com/acme/relay/internal/hub.(*Hub).Run()\n /src/internal/hub/hub.go:84 +0xd8\n==================\nFAIL github.com/acme/relay/internal/hub 3.412s\nok github.com/acme/relay/internal/api 1.204s\nFAIL\nmake: *** [Makefile:31: test] Error 1\n\nthis passed on the PR branch with the same commit sha, which makes no sense to me", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "pasted-context", "lang": "en"}
|
|
{"prompt": "the model's eval accuracy dropped 6 points after we changed nothing but the data loader. same seed, same checkpoint", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
|
{"prompt": "keyboard pushes the input off screen on some androids and not others. no idea", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
|
{"prompt": "look over the new caching layer and tell me where the invalidation can go wrong. i'm not aware of a bug, i just want a second pair of eyes before we roll it out", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.65, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "how does our request id propagate from the edge worker through to the db slow query log", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
|
{"prompt": "worth switching from bcrypt to argon2id for us? we have 900k users and a hot login path", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.55, "slice": "core", "lang": "en"}
|
|
{"prompt": "the auth module has zero comments and three non-obvious invariants. document them inline where they matter, and add a short doc/auth-invariants.md", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "ticket text for the backlog item we talked about in standup — 'inventory counts drift after partial refunds'. i want repro steps, impact, and acceptance criteria written properly", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.35, "slice": "core", "lang": "en"}
|
|
{"prompt": "escreve um resumo técnico da arquitetura atual do serviço de pagamentos para o time novo, mais ou menos uma página", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "pt"}
|
|
{"prompt": "fix the sort order thing", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "vague-eval", "lang": "en"}
|
|
{"prompt": "picking the search relevance work back up. plan the reranking approach and then implement the first pass — we have bm25 from opensearch and 1.2M docs with click logs", "purpose": "planning", "secondary": "backendImpl", "mixed": true, "difficulty": 0.85, "slice": "mixed", "lang": "en"}
|
|
{"prompt": "reduce our aws bill. i have the cost explorer breakdown, biggest lines are nat gateway data processing, rds on-demand, and cloudwatch logs ingestion. give me a prioritized plan", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.75, "slice": "core", "lang": "en"}
|
|
{"prompt": "seed script that builds a realistic dev dataset: 50 orgs, skewed user counts, a year of activity with weekday/weekend shape, deterministic from a seed arg", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
|
{"prompt": "toast notifications, stacked, max 3 visible, swipe to dismiss, and they shouldn't cover the FAB", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
|
{"prompt": "shave the extra pixel off the divider under the nav", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.05, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "extract the rate limit config into one place, it's currently hardcoded in four middleware files with different numbers", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.45, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "double charge happened for 6 customers on the 14th. i need to know how before i tell finance", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.85, "slice": "core", "lang": "en"}
|
|
{"prompt": "is the way we're doing optimistic UI in the comment box actually safe if the mutation fails mid-flight? no bug reported, just want your read on it", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.55, "slice": "boundary", "lang": "en"}
|
|
{"prompt": "write up how the deploy pipeline works for the wiki, from merge to prod, including the manual approval gate and who can click it", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.45, "slice": "core", "lang": "en"}
|
|
{"prompt": "just make it work", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "vague-eval", "lang": "en"}
|
|
{"prompt": "yesterday's grafana dashboard work — finish wiring the panels", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "vague-eval", "lang": "en"}
|