adding purpose-classifier data
This commit is contained in:
@@ -0,0 +1,200 @@
|
||||
{"prompt": "we need to get the notification fan-out off the primary postgres. write me an RFC comparing SNS+SQS, a self-hosted NATS cluster, and just using pg_notify better — include a rollout order and where the kill switches go", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
||||
{"prompt": "design the multi-tenant audit log schema for us — 7 year retention, queryable by actor and by resource, tenants can export their own slice. i only want the design doc right now, don't touch the migrations", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "plan and then start building the offline write queue for the android client. ordering guarantees matter, we can't have a note edit land before its create", "purpose": "planning", "secondary": "backendImpl", "mixed": true, "difficulty": 0.8, "slice": "mixed", "lang": "en"}
|
||||
{"prompt": "add a /v2/exports endpoint to the go service that streams a zipped csv per tenant — chunked writes, never buffer the whole archive in memory, and it has to survive a 20 minute download", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "idempotency keys on POST /charges please. 24h window, keys in redis, replay returns the original response body and status verbatim", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "from the notion doc, please build it:\n\n- POST /v1/holds { seat_ids: [uuid], hold_ttl_seconds: int (default 600, max 1800) }\n- returns 201 { hold_id, expires_at } or 409 with the conflicting seat ids\n- holds expire without a cron — lazy expiry on read is fine\n- a second hold on the same seats from the SAME session extends instead of conflicting\n- DELETE /v1/holds/:id releases early, must be idempotent\n- all of it behind the seat_holds flag\n\npostgres + sqlc, copy the repo layer patterns from internal/reserve", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "pasted-context", "lang": "en"}
|
||||
{"prompt": "build the settings screen in SwiftUI — sections for account, notifications, appearance. use Form, match the row styling in ProfileView, and the appearance section needs a live preview tile", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "dark mode toggle for the docs site", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
||||
{"prompt": "swap the spinner on the feed for the shimmer placeholder we already use in Notifications", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "bump ruff to 0.6.9 in pyproject and clear whatever new lint errors it throws", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "core", "lang": "en"}
|
||||
{"prompt": "typo: 'unathorized' in auth/middleware.go", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "core", "lang": "en"}
|
||||
{"prompt": "one line thing — the retry cap in worker.rs should be 5 not 3. double check nothing else reads that const before you change it", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "PaymentsCoordinator is 1400 lines and i can't test any of it. split the state machine out from the networking, identical behavior, the existing snapshot tests should still pass untouched", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "the retry + backoff logic is copy pasted across all six of our http clients. pull it into one middleware and have them all use it", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "rename the `usr` param to `user` in handlers/*.go, it's like 40 call sites but purely mechanical", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "our nightly ETL hangs about 40 minutes in. no error, no cpu, no log lines, just sits there until the k8s deadline kills it. maybe 1 run in 4. figure out what's going on", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
||||
{"prompt": "got this on the payments worker an hour ago, third time this week\n\npanic: runtime error: invalid memory address or nil pointer dereference\n[signal SIGSEGV: segmentation violation code=0x1 addr=0x0 pc=0x8f2a41]\n\ngoroutine 4417 [running]:\ngithub.com/acme/ledger/internal/settle.(*Batcher).flush(0xc000a3e000, {0x1043a20, 0xc0004b2340})\n\t/app/internal/settle/batcher.go:212 +0x1c1\ngithub.com/acme/ledger/internal/settle.(*Batcher).run(0xc000a3e000)\n\t/app/internal/settle/batcher.go:141 +0xf5\ncreated by github.com/acme/ledger/internal/settle.NewBatcher in goroutine 1\n\t/app/internal/settle/batcher.go:88 +0x145\n\nbatcher.go hasn't been touched since march. i'm lost", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "pasted-context", "lang": "en"}
|
||||
{"prompt": "why is my build 4x slower since friday", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "walk me through the refill math in internal/ratelimit, specifically why there's a second clamp after the burst add", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
||||
{"prompt": "if i drop the FOR UPDATE in reserveSeat does that open a race, or does the unique constraint already save us? just want your read, don't change anything yet", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "go through the terraform in infra/ and tell me if anything lets a dev-role principal touch prod state or assume the deploy role", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "release notes for 3.4.0 off the commits since 3.3.2, split into user-facing and internal, and call out the deprecation", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "core", "lang": "en"}
|
||||
{"prompt": "the openapi spec still says amount is an integer in cents, we switched it to a decimal string last month. get the spec and the endpoint descriptions current", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "write the cutover runbook for the shard migration and also add the pre-flight check script it references", "purpose": "writing", "secondary": "backendImpl", "mixed": true, "difficulty": 0.6, "slice": "mixed", "lang": "en"}
|
||||
{"prompt": "continue", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "vague-eval", "lang": "en"}
|
||||
{"prompt": "we want the feature store off pandas-in-airflow and onto something real. i have three engineers and one quarter. give me a roadmap with what we cut", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.9, "slice": "core", "lang": "en"}
|
||||
{"prompt": "thinking out loud here — every service reinvents pagination, some cursor some offset, the mobile team is furious. i want one story for pagination across all of them but i don't know if we version the APIs or add a parallel param set or what. help me think it through and land on a strategy", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
||||
{"prompt": "dumping the migration TODOs from our doc, turn it into a sequenced plan with dependencies called out, i suspect half of these can't happen in the order listed:\n\n- move sessions table to its own database\n- kill the legacy /api/session endpoint (mobile 4.2 still calls it, 11% of DAU)\n- switch session ids from int to uuidv7\n- stop writing session rows on anonymous traffic\n- add the read replica for analytics\n- delete the sessions_archive_2019 table (nobody knows if finance reads it)\n- rotate the signing key, needs dual-accept window\n- redis session cache TTL down from 7d to 24h", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "pasted-context", "lang": "en"}
|
||||
{"prompt": "Room migration for the offline notes table — add synced_at, backfill it from updated_at, and don't destructive-migrate, people have unsynced data", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "duckdb loader for the parquet exports. partition by dt, skip anything already in the manifest table, and it should be safe to run twice", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "the api docs already describe GET /v1/segments in full detail but the endpoint doesn't exist. make the code match the docs", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "compose bottom sheet for the filter chips, sticky Apply button at the bottom, material3 tokens only, and it should keep its selection across rotation", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "make the chart legend wrap", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "core", "lang": "en"}
|
||||
{"prompt": "priya left this in #design-review, can you do it:\n\n\"ok looked at the build on my 13 mini. the header eats like a third of the screen — the avatar and the greeting can be one row not two. the segmented control is fine but the selected pill has no contrast in dark mode, it's basically invisible. the card shadows are way too heavy, we said 4% not 20%. and the pull to refresh spinner appears BEHIND the header which looks broken. everything else 👍\"", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "pasted-context", "lang": "en"}
|
||||
{"prompt": "gradle targetSdk to 35", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "core", "lang": "en"}
|
||||
{"prompt": "the s3 lifecycle rule should expire at 90 days not 30, it's in infra/storage somewhere", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "core", "lang": "en"}
|
||||
{"prompt": "alembic's env.py has compare_type=False, flip it on and regenerate nothing", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "the feature engineering lives in notebooks/features.ipynb and gets copy-pasted into prod by hand. lift it into a real module, same numbers out, add tests that pin the current output", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "FeedViewModel does retrofit calls directly in four places. push all of it behind FeedRepository like the other screens do", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "we have two parse_timestamp helpers that disagree on trailing Z. collapse them into one, keep the utils/time.py behavior since more callers depend on it", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "duplicate rows show up in the daily rollup but only when a backfill window overlaps a live run. the merge key looks right to me. i don't get it", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "ANR in FeedFragment after rotation, ~60% of the reports are pixels on android 14. logcat isn't telling me much", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "this passes locally every time:\n\n=================================== FAILURES ===================================\n_______________________ test_rollup_dedupes_late_events ________________________\n\n def test_rollup_dedupes_late_events():\n run_rollup(day=date(2026, 3, 14))\n> assert warehouse.count(\"rollup_daily\", day=\"2026-03-14\") == 1\nE assert 2 == 1\nE + where 2 = count('rollup_daily', day='2026-03-14')\n\ntests/test_rollup.py:118: AssertionError\n----------------------------- Captured log call ------------------------------\nWARNING rollup.merge:merge.py:64 late event 4a91 arrived after watermark 2026-03-14T23:59:59Z\nINFO rollup.merge:merge.py:71 upsert conflict target (day, tenant) -> 1 row\nINFO rollup.merge:merge.py:71 upsert conflict target (day, tenant) -> 1 row\n=========== 1 failed, 214 passed, 3 skipped in 41.22s ===========", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "pasted-context", "lang": "en"}
|
||||
{"prompt": "after we added the partial index the planner switched to a seq scan on the orders join. expected or did we break something", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "explain what our celery beat schedule actually does, the crontab entries are impossible to read and i don't trust the comments", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
||||
{"prompt": "帮我看下这个 PR 里的 SQL migration 有没有锁表风险,表大概 2 亿行。只要分析,先别改代码", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "zh"}
|
||||
{"prompt": "README for the dbt project — how to run it locally, how to add a model, the naming conventions we settled on, and the bit about not committing target/", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
||||
{"prompt": "pr description for this branch. lead with the breaking change to the webhook payload shape so reviewers can't miss it", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "core", "lang": "en"}
|
||||
{"prompt": "write up the postmortem for last night's outage and then land the missing index that caused it", "purpose": "writing", "secondary": "quickFix", "mixed": true, "difficulty": 0.5, "slice": "mixed", "lang": "en"}
|
||||
{"prompt": "picking the feature store thing back up", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "vague-eval", "lang": "en"}
|
||||
{"prompt": "architecture for the on-device inference path: we want the tiny classifier to run in a coreml model but fall back to the server when the device is old. where does the decision live, how do we keep both versions in sync, what do we log", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
||||
{"prompt": "milestones for killing our custom feature flag service and moving to openfeature. weekly chunks, each one revertible on its own", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "figure out an approach for the firmware OTA rollout — staged by device cohort, must survive a power cut mid-flash, and we only have 128k for the bootloader. write it up before any code", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.9, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "rust: parse the .ndjson event feed, tolerate trailing garbage on a line, emit a counter for malformed records instead of bailing out", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "nightly celery job that reconciles stripe invoices against our ledger, idempotent, writes discrepancies to a table instead of alerting", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "postgres migration: composite index on events (tenant_id, created_at desc), concurrently, table is 800M rows and we can't take a lock", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "skeleton loading state for the dashboard cards. shimmer, but bail to a static grey if prefers-reduced-motion", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
||||
{"prompt": "animate the tab bar selection with matchedGeometryEffect, and the icon should scale a touch on tap", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "sidebar collapses at 1024 but tablet users get a mangled layout around 900. redo the breakpoints, and the nav should stay reachable at every width", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "docker healthcheck interval to 10s", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "core", "lang": "en"}
|
||||
{"prompt": "in der nginx.conf client_max_body_size auf 50m setzen, uploads schlagen bei 12mb fehl", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "core", "lang": "de"}
|
||||
{"prompt": "new_pricing_page on in staging only, prod stays off", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "core", "lang": "en"}
|
||||
{"prompt": "three copies of formatCurrency — web, mobile-web, admin. one implementation, one package, delete the others", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "convert the callback-style s3 helpers in lib/storage to async/await. keep the exported signatures compatible where you can and note anywhere you couldn't", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "SearchController knows about the network layer, the cache, analytics AND the keyboard. decouple it — protocols at the boundaries, no behavior drift, the ui should look identical", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "the websocket gateway grows about 200mb an hour and gets oomkilled overnight. node 22. heap snapshots look boring, nothing obviously retained", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
||||
{"prompt": "app freezes on launch, iOS 17 only, 18 is fine. main thread checker yells something about CoreData and then nothing", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "sentry issue, 340 events in 2h, all from the same customer:\n\nTypeError: Cannot read properties of undefined (reading 'currency')\n at formatTotal (checkout/summary.tsx:88:31)\n at CartSummary (checkout/summary.tsx:142:18)\n at renderWithHooks (react-dom.production.min.js:1:1)\n\nbreadcrumbs\n navigation /cart -> /checkout\n xhr GET /api/cart 200 (412ms)\n console warn: cart line 9f21 has no price snapshot\n xhr POST /api/checkout/quote 500 (88ms)\n ui.click button[data-test=place-order]\n\ntags: release 2026.7.3, browser chrome 141, locale de-CH\n\nour cart response is supposed to always have currency. what's actually happening", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "pasted-context", "lang": "en"}
|
||||
{"prompt": "is our jwt refresh flow safe if someone lifts a refresh token out of localStorage? rotation is on but i want a second pair of eyes on the reuse detection", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "what does this regex do", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "core", "lang": "en"}
|
||||
{"prompt": "bull vs temporal for our job layer — read what we have in workers/ and tell me honestly whether switching buys us anything. no code", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "godoc comments on everything exported in pkg/graph. mention the panic conditions, they're not obvious", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
||||
{"prompt": "changelog entry: we removed --legacy-auth from the cli, and there's no replacement flag, people need to use the config file", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "core", "lang": "en"}
|
||||
{"prompt": "commit message for what's staged", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "core", "lang": "en"}
|
||||
{"prompt": "map out how we'd add row level security to the reporting db without breaking the 40-odd existing queries, then do the first table as a proof", "purpose": "planning", "secondary": "backendImpl", "mixed": true, "difficulty": 0.8, "slice": "mixed", "lang": "en"}
|
||||
{"prompt": "clean up whatever's left", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "vague-eval", "lang": "en"}
|
||||
{"prompt": "we're getting rate limited by the maps provider at peak. write up options — client side caching, a proxy with a shared cache, or negotiating tiers — with cost estimates and what each one costs us in latency", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "i want a design for replay: given an aggregate id we should be able to rebuild its state from the event log at any point in time. we have 14 event types and three of them changed shape without versioning. how do we handle that", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.9, "slice": "core", "lang": "en"}
|
||||
{"prompt": "grpc streaming endpoint for live telemetry, server-side only, 5k concurrent devices, backpressure via a bounded channel and drop-oldest", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "cli subcommand `vault export --format=env` that dumps decrypted secrets to stdout, refuses if stdout is a tty unless you pass --force", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "implement the sliding window counter for the abuse limiter in go, per-ip and per-account, redis lua so it's atomic", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "the empty state for Saved Items is just the word 'empty'. give it an illustration slot, a headline, a subhead and a CTA that deep links to browse", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
||||
{"prompt": "toast notifications — stack up to 3, auto dismiss at 4s, pause the timer on hover, swipe to dismiss on touch. no libraries, we already have framer motion", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "kerning on the pricing table headers is off and the third column is 2px wider than the others for no reason i can find", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "pin actions/checkout to a sha instead of v4 in all the workflows", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "log line says 'Retrying in {}s' with no arg filled in, fix the format call", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "core", "lang": "en"}
|
||||
{"prompt": "clippy is unhappy after the toolchain bump:\n\nwarning: this expression creates a reference which is immediately dereferenced by the compiler\n --> src/index/segment.rs:214:27\n |\n214 | self.merge_into(&mut writer, &fields)\n | ^^^^^^^ help: change this to: `fields`\n = note: `-W clippy::needless-borrow` implied by `-D warnings`\n\nwarning: called `map(..).flatten()` on an `Iterator`\n --> src/index/segment.rs:301:18\n |\n301 | .map(|d| d.terms()).flatten()\n | ^^^^^^^^^^^^^^^^^^^^^^^^^^^^ help: try: `flat_map(|d| d.terms())`\n\nwarning: redundant clone\n --> src/query/plan.rs:77:41\n |\n77 | let filters = self.filters.clone();\n | ^^^^^^^^ help: remove this\n\nerror: could not compile `tantivy-lite` (lib) due to 3 warnings\n\nCI is red on -D warnings, just get it green", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "pasted-context", "lang": "en"}
|
||||
{"prompt": "kill the singleton in AnalyticsManager.shared, inject it instead. every call site, tests included, same events firing at the same times", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "hoist the duplicated zod schemas out of the route files into schemas/ and re-export, nothing about validation behavior should change", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "swap our hand rolled Result type for anyhow across the crate. it's mechanical but there are a few places we match on the error variant", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "test_scheduler.py::test_backoff_jitter fails maybe 5% of the time in CI, never once locally, and the assertion is on a duration so i assume timing but i've been wrong about that before", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "pods keep restarting after the deploy, here's kubectl describe:\n\nEvents:\n Type Reason Age From Message\n ---- ------ ---- ---- -------\n Normal Pulled 8m kubelet Container image \"registry/api:2026.7.4\" already present\n Normal Created 8m kubelet Created container api\n Normal Started 8m kubelet Started container api\n Warning Unhealthy 7m (x3 over 7m) kubelet Readiness probe failed: HTTP probe failed with statuscode: 503\n Warning Unhealthy 6m (x2 over 6m) kubelet Liveness probe failed: Get \"http://10.42.3.19:8080/healthz\": context deadline exceeded\n Normal Killing 6m kubelet Container api failed liveness probe, will be restarted\n Warning BackOff 2m (x14 over 6m) kubelet Back-off restarting failed container\n\nthe same image runs fine in staging with the same manifests", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "pasted-context", "lang": "en"}
|
||||
{"prompt": "warum liefert unser Cache nach dem Deploy gelegentlich Daten vom Vortag? TTL ist 60s, ich verstehe es nicht", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "de"}
|
||||
{"prompt": "read through migrations/ and tell me which of these are not reversible, i want to know before we promise a rollback plan", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "second opinion on this hunk before i merge:\n\n@@ -41,15 +41,22 @@ func (s *Store) Claim(ctx context.Context, workerID string) (*Job, error) {\n-\trow := s.db.QueryRowContext(ctx, `\n-\t\tUPDATE jobs SET state='running', worker_id=$1, claimed_at=now()\n-\t\tWHERE id = (SELECT id FROM jobs WHERE state='queued' ORDER BY priority DESC, id LIMIT 1)\n-\t\tRETURNING id, payload`, workerID)\n+\ttx, err := s.db.BeginTx(ctx, nil)\n+\tif err != nil {\n+\t\treturn nil, err\n+\t}\n+\tdefer tx.Rollback()\n+\trow := tx.QueryRowContext(ctx, `\n+\t\tSELECT id, payload FROM jobs\n+\t\tWHERE state='queued' ORDER BY priority DESC, id\n+\t\tLIMIT 1 FOR UPDATE SKIP LOCKED`)\n \tvar j Job\n \tif err := row.Scan(&j.ID, &j.Payload); err != nil {\n \t\treturn nil, err\n \t}\n+\tif _, err := tx.ExecContext(ctx, `UPDATE jobs SET state='running', worker_id=$1, claimed_at=now() WHERE id=$2`, workerID, j.ID); err != nil {\n+\t\treturn nil, err\n+\t}\n+\treturn &j, tx.Commit()\n }", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "pasted-context", "lang": "en"}
|
||||
{"prompt": "is it actually a problem that we hold the mutex across the http call in fetchProfile, or am i being paranoid", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "migration guide for people on v2 of the sdk. the auth constructor changed, retries are opt-out now instead of opt-in, and the pagination helper is gone", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "translate the error message catalog in i18n/errors.en.json to fr and pt-BR, keep the placeholders exactly as they are", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
||||
{"prompt": "the /webhooks docs page doesn't mention retry semantics or the signature header at all. document both, and be precise about the timestamp tolerance window", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "do the thing we talked about with the queue", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "vague-eval", "lang": "en"}
|
||||
{"prompt": "shape up a proposal for how we test the embedded firmware in CI. we have 6 boards on a bench and no automation. include what we fake vs what we run on metal", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
||||
{"prompt": "propose the data model for a shared shopping list — offline edits on multiple devices, we need last-writer-wins per field not per row, and deletes have to survive a sync that arrives out of order", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.9, "slice": "core", "lang": "en"}
|
||||
{"prompt": "estrategia para migrar del monolito Django a servicios, tenemos 8 meses. quiero el orden de extracción y los riesgos, sin código todavía", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.9, "slice": "core", "lang": "es"}
|
||||
{"prompt": "webhook delivery service: at-least-once, exponential backoff to 24h, dead letter after that, per-endpoint concurrency cap so one slow customer can't starve the pool", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "add SAML SSO to the admin panel, okta first. just-in-time provisioning, and the group claim maps to our roles table", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
||||
{"prompt": "write the tree-sitter based symbol indexer for the language server, incremental on edit, python and ts to start", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
||||
{"prompt": "drag to reorder for the playlist rows in SwiftUI, haptic on pickup, and the row being dragged should lift with a shadow", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "the game's pause menu needs controller focus handling — dpad moves selection, B backs out, and the currently focused item should be obvious at a glance on a tv", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "make the copy button in the code blocks actually visible instead of appearing on hover only, mobile users can't get to it", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "vite config, raise chunkSizeWarningLimit, the warning is noise", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "core", "lang": "en"}
|
||||
{"prompt": "we spell it 'canceled' in half the ui strings and 'cancelled' in the other half. pick the double-l one and make it uniform", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "the cors config is missing the new preview domain, add *.preview.acme.dev", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "core", "lang": "en"}
|
||||
{"prompt": "flatten the three levels of inheritance under BaseRepository into composition. i don't want to change a single query", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "split the god module utils.py — 2200 lines, everything imports it, circular imports everywhere. by concern, keep re-exports so nothing breaks", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "rename the whole `Widget` concept to `Panel` — types, files, css classes, storybook stories, the graphql field too. it's the term everyone actually uses", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "the mac app hangs for ~8 seconds on first launch after an update, spindump shows a bunch of threads waiting on something in our keychain code but i can't tell what's blocking", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
||||
{"prompt": "search results come back in a different order for the same query depending on which replica answers. shouldn't the sort be deterministic", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "asan on the firmware sim:\n\n==2917==ERROR: AddressSanitizer: heap-buffer-overflow on address 0x60300000ef34 at pc 0x000104a1b2c1 bp 0x7ff7bfef1f10 sp 0x7ff7bfef1ed0\nREAD of size 4 at 0x60300000ef34 thread T0\n #0 0x104a1b2c0 in ring_pop ring.c:88\n #1 0x104a19a44 in uart_drain uart.c:213\n #2 0x104a1c110 in main_loop main.c:64\n\n0x60300000ef34 is located 0 bytes to the right of 20-byte region [0x60300000ef20,0x60300000ef34)\nallocated by thread T0 here:\n #0 0x10a4c1b90 in wrap_calloc\n #1 0x104a1ae08 in ring_init ring.c:31\n\nSUMMARY: AddressSanitizer: heap-buffer-overflow ring.c:88 in ring_pop\n\nonly trips when the buffer wraps. i've stared at the modulo for an hour", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "pasted-context", "lang": "en"}
|
||||
{"prompt": "explain the ownership model in src/session — who's allowed to mutate a Session and when, because i keep getting borrow errors and i think i'm holding it wrong conceptually", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "how does our feature flag evaluation order work when a user matches two segments", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
||||
{"prompt": "give me a tour of the build pipeline in Makefile + scripts/, i inherited this repo yesterday", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "adr for the decision to keep sessions in postgres instead of redis. context, options we looked at, consequences, the usual shape", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "summarize what changed in the auth service over the last two weeks for the team update, plain language, no commit hashes", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "onboarding doc for new backend hires — local setup, how to get seed data, the four services they'll actually touch, who to ask about what", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "outline how we'd add end to end encryption to attachments and then wire up the key derivation piece", "purpose": "planning", "secondary": "backendImpl", "mixed": true, "difficulty": 0.9, "slice": "mixed", "lang": "en"}
|
||||
{"prompt": "fix the flaky one and write the note for the ticket about why it was flaky", "purpose": "debugging", "secondary": "writing", "mixed": true, "difficulty": 0.6, "slice": "mixed", "lang": "en"}
|
||||
{"prompt": "make it pop", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "vague-eval", "lang": "en"}
|
||||
{"prompt": "i want to introduce a read model for the inventory dashboard because the joins are killing us. write the design — where projections live, how we rebuild, what staleness we accept — and flag anything that forces an api change", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
||||
{"prompt": "how should we structure the monorepo now that we have 3 apps and 11 packages? turborepo is in already but the boundaries are mush. give me a target layout and a path to get there", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "batch import endpoint, csv up to 500mb, presigned upload then async processing, per-row errors reported back with line numbers", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "full text search over the notes table with tsvector, weighted title over body, and a trigger to keep the column current", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "the docs promise an `X-Request-Id` echo on every response but the middleware only sets it on 2xx. make reality match the docs", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "onboarding carousel, 4 pages, page dots, skip in the corner, and it should never show again after completion", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "virtualize the log viewer, we render 200k rows and the tab dies. keep the sticky timestamp column and the find-in-page highlight working", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "spacing on the invoice pdf preview is off — line items too tight, totals block should hug the right edge with a rule above it", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
||||
{"prompt": "raise the pg pool from 10 to 25 in config/database.yml", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "core", "lang": "en"}
|
||||
{"prompt": "npm audit is screaming about the tar transitive dep, get us onto a patched version without breaking the build", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "core", "lang": "en"}
|
||||
{"prompt": "the date format string is MM/DD/YYYY in ExportPanel but everywhere else we use ISO, change it there too", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "untangle the circular dep between the auth package and the user package. probably an interface in a third place, behavior stays put", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "our reducers all have the same 30 lines of optimistic-update bookkeeping. factor it into one helper and adopt it in all seven", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "simplify these nested ternaries in PriceBadge, they're unreadable. same rendered output, i'll diff the snapshots", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "requests to the payments provider started timing out at exactly 10s about 40% of the time, started tuesday, nothing deployed tuesday. their status page is green", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
||||
{"prompt": "tail of the failing CI job, this is the fourth rerun:\n\n ✓ apps/web:test (48.2s)\n ✓ packages/ui:test (12.9s)\n ✗ apps/api:test\n\n FAIL src/queue/consumer.test.ts > consumer > acks after successful handler\n Error: Timeout of 5000ms exceeded. The last call to done() was never made.\n at listOnTimeout (node:internal/timers:581:17)\n\n stderr | src/queue/consumer.test.ts\n [amqp] connection closed unexpectedly (code 320, CONNECTION_FORCED)\n [amqp] reconnect attempt 1 in 500ms\n [amqp] reconnect attempt 2 in 1000ms\n\n Test Files 1 failed | 22 passed (23)\n Tests 1 failed | 318 passed (319)\n Duration 91.44s\n\nERROR: command finished with error: exit 1\n\nit passed 60 times in a row before this week", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "pasted-context", "lang": "en"}
|
||||
{"prompt": "iOS widget shows stale data until you open the app, timeline provider looks right to me", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "read the caching layer in internal/cache and tell me under what conditions we can serve a stale value. i'm not saying anything's wrong, i just need to know the guarantees", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "security pass on the file upload path — content type sniffing, path traversal in the key, and whether a signed url can be replayed", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "où est-ce que le token est validé exactement dans le middleware? explique moi le chemin complet d'une requête authentifiée", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "fr"}
|
||||
{"prompt": "blog post draft about how we cut p95 from 900ms to 210ms. technical audience, real numbers, no marketing voice", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "inline comments in scheduler.go explaining the priority inversion guard, future me will not remember why that sleep is there", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "deprecation notice for the /v1 endpoints — sunset date, what to use instead, and the header we'll start returning", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "core", "lang": "en"}
|
||||
{"prompt": "just get it working", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "vague-eval", "lang": "en"}
|
||||
{"prompt": "we keep hitting the 6MB lambda response limit on the report endpoint. lay out the options — s3 handoff, pagination, streaming through an alb — and recommend one given our client is a react app that renders the whole table", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "break the accessibility work into shippable chunks. audit found 140 issues across 30 screens and leadership wants a quarterly commitment", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "設計だけお願いしたい。マルチテナントの権限モデルを RBAC から ReBAC に寄せたい場合の移行方針を書いて、既存の 60 個くらいのパーミッションチェックへの影響も洗い出して", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.9, "slice": "core", "lang": "ja"}
|
||||
{"prompt": "wire up outbox pattern for the order events — write to outbox in the same tx, a poller publishes to kafka, dedupe on the consumer side by event id", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "python service that pulls the hourly weather grib files, subsets to our bounding boxes and writes zarr. retries on partial downloads, resumable", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "add cursor pagination to GET /activity. opaque cursor, stable under inserts, include has_more, and don't break the existing page/limit callers", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "keyboard shortcut palette, cmd+k, fuzzy filter, recent commands at top, and it has to trap focus properly", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "the map pins overlap at low zoom, cluster them and animate the expand when you zoom in", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "shrink the avatar in the comment header, it's competing with the name", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "prod log level is debug, drop it to info before we burn more money", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "core", "lang": "en"}
|
||||
{"prompt": "off by one in the pagination footer, it says 'showing 1-20 of 20' when there are 21", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "core", "lang": "en"}
|
||||
{"prompt": "esse timeout de 3s no client HTTP tá curto demais, sobe pra 15s e deixa configurável por env", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "core", "lang": "pt"}
|
||||
{"prompt": "move the 12 test helpers out of conftest.py into a fixtures package, group by domain, all 900 tests should pass unchanged", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "replace our bespoke event emitter with node's EventEmitter, keep the same subscribe/unsubscribe semantics including the once behavior", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "extract the 200 line switch in handleCommand into a dispatch table, one file per command, no functional change", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "csv exports come out with mojibake for anything non-ascii but only when downloaded from safari. same bytes look fine in curl", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "tsc suddenly unhappy after pulling main, i didn't touch types:\n\nsrc/features/billing/useSubscription.ts:44:5 - error TS2322: Type '{ status: string; renewsAt: string | null; }' is not assignable to type 'Subscription'.\n Types of property 'status' are incompatible.\n Type 'string' is not assignable to type 'SubscriptionStatus'.\n\n44 return { status: data.status, renewsAt: data.renews_at }\n ~~~~~~\n\nsrc/features/billing/PlanCard.tsx:19:34 - error TS2339: Property 'renewsAt' does not exist on type 'Subscription | LegacySubscription'.\n Property 'renewsAt' does not exist on type 'LegacySubscription'.\n\n19 const label = formatRenewal(sub.renewsAt)\n ~~~~~~~~\n\nFound 2 errors in 2 files.\n\nnobody on the team knows where LegacySubscription came from", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "pasted-context", "lang": "en"}
|
||||
{"prompt": "gpu memory climbs across training epochs until it OOMs around epoch 30. torch 2.6, i'm detaching the loss, i think", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
||||
{"prompt": "does our retry logic in the sms client risk double-sending? walk me through what happens on a timeout after the provider already accepted", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "review the new dockerfile, i'm mostly worried about image size and whether we're leaking build secrets into a layer", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "what's the difference between our two queue implementations and why do both exist", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "user facing help article for the new saved-views feature. friendly tone, screenshots marked as TODO where they go", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "core", "lang": "en"}
|
||||
{"prompt": "the CONTRIBUTING file predates the monorepo move and tells people to run yarn. rewrite it for pnpm and the new package layout", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "core", "lang": "en"}
|
||||
{"prompt": "rfc doc for the rewrite plus a proof of concept of the new renderer so people can see it's not vapor", "purpose": "planning", "secondary": "frontendImpl", "mixed": true, "difficulty": 0.9, "slice": "mixed", "lang": "en"}
|
||||
{"prompt": "explain what the throttle in useScrollSpy is doing and then rewrite it with the shared hook", "purpose": "review", "secondary": "refactor", "mixed": true, "difficulty": 0.4, "slice": "mixed", "lang": "en"}
|
||||
{"prompt": "back to the payment flow bug from yesterday", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "vague-eval", "lang": "en"}
|
||||
{"prompt": "we need a story for schema evolution on the kafka topics. avro + registry, or protobuf, or just json and pray. write it up with what each costs us operationally", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
||||
{"prompt": "sketch the state machine for the checkout flow. states, allowed transitions, what happens on a browser back, where we persist. diagram in mermaid is fine", "purpose": "planning", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "rate limiter for the public api — tiered by plan, burst allowance, 429 with retry-after, and the counters can't be a single redis hot key", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "soft delete on the projects table plus the scoping so nothing leaks into list endpoints, and a hard-purge job for the 30 day mark", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "go worker that watches the s3 bucket via sqs notifications and thumbnails anything new, three sizes, skip if the derivative already exists", "purpose": "backendImpl", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "inline editing on the table cells — click to edit, esc cancels, enter commits and moves down a row, optimistic with a rollback on error", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "hero section for the landing page: big headline, gradient mesh behind it, two ctas, and it can't jank on scroll", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.5, "slice": "core", "lang": "en"}
|
||||
{"prompt": "the destructive action in the context menu should be red and last in the list, right now it's first and looks like everything else", "purpose": "frontendImpl", "secondary": null, "mixed": false, "difficulty": 0.2, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "add a .nvmrc with 22", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "core", "lang": "en"}
|
||||
{"prompt": "the ignore list in .gitignore is missing .env.local and someone will commit it eventually", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.1, "slice": "core", "lang": "en"}
|
||||
{"prompt": "our health endpoint returns 200 even when the db check fails, it should be 503. small change in health.go, i know exactly where", "purpose": "quickFix", "secondary": null, "mixed": false, "difficulty": 0.3, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "the graphql resolvers reach into the ORM directly. put a service layer between them, one service per domain, resolvers get thin. no schema changes", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "swift concurrency cleanup in NetworkClient — it's half completion handlers half async, unify on async and mark the actor boundaries properly", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "core", "lang": "en"}
|
||||
{"prompt": "kill the dead code under legacy/ — but check nothing outside legacy/ imports it first, including the cron scripts", "purpose": "refactor", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "two users report their calendar events shifted by an hour but only for events created before the DST switch. we store utc so i genuinely don't see how", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
||||
{"prompt": "docker build dies on the arm runner only:\n\n #12 [builder 6/9] RUN cargo build --release --locked\n #12 24.31 Compiling ring v0.17.8\n #12 31.02 error: failed to run custom build command for `ring v0.17.8`\n #12 31.02 \n #12 31.02 Caused by:\n #12 31.02 process didn't exit successfully: `/app/target/release/build/ring-8f1d/build-script-build` (exit status: 101)\n #12 31.02 --- stderr\n #12 31.02 thread 'main' panicked at build.rs:684:9:\n #12 31.02 error building ring: exit status: 1\n #12 31.02 note: run with `RUST_BACKTRACE=1`\n------\nfailed to solve: process \"/bin/sh -c cargo build --release --locked\" did not complete successfully: exit code: 101\n\nx86 builds fine, same dockerfile, same lockfile", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.7, "slice": "pasted-context", "lang": "en"}
|
||||
{"prompt": "intermittent 502s from the ingress, maybe 1 in 500 requests, no matching error in the app logs at all", "purpose": "debugging", "secondary": null, "mixed": false, "difficulty": 0.8, "slice": "core", "lang": "en"}
|
||||
{"prompt": "does the new websocket auth actually prevent a user from subscribing to another tenant's channel? show me where that's enforced", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "read AGENTS.md and the two scripts it references and tell me if a fresh clone could actually follow it", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
||||
{"prompt": "how much of the query cost in the dashboard comes from the permission filter vs the actual data fetch, based on reading the code not profiling", "purpose": "review", "secondary": null, "mixed": false, "difficulty": 0.6, "slice": "core", "lang": "en"}
|
||||
{"prompt": "terraform docs for the vpc module — inputs, outputs, an example, and the note about why the nat gateway is single-az on purpose", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "boundary", "lang": "en"}
|
||||
{"prompt": "changelog for the last three releases, we never wrote them and now someone asked", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
||||
{"prompt": "docstrings for the pandas transforms in etl/transform.py, including what each one assumes about the input index", "purpose": "writing", "secondary": null, "mixed": false, "difficulty": 0.4, "slice": "core", "lang": "en"}
|
||||
Reference in New Issue
Block a user