adding purpose-classifier data
This commit is contained in:
@@ -0,0 +1,227 @@
|
||||
{"prompt":"Design a migration from hand-authored feature SQL to a governed feature platform. Cover ownership, point-in-time correctness, offline and online parity, lineage, schema evolution, backfills, validation gates, cost controls, adoption milestones, and rollback for models that regress after switching sources.","purpose":"planning","secondary":null,"mixed":false,"difficulty":0.9,"slice":"core","lang":"en"}
|
||||
{"prompt":"Implement a Python worker that joins click events to the latest eligible catalog snapshot without leaking future data.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Build the model-run comparison page with aligned metric charts, dataset filters, confidence intervals, and shareable URLs.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Bump scikit-learn to 1.7.1","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"core","lang":"en"}
|
||||
{"prompt":"Extract common tensor validation from the three inference adapters without changing accepted shapes or error messages.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"Our recommender returns nearly identical results for every user after several hours. Find where embeddings or cache keys collapse.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Audit the training split code for temporal leakage","purpose":"review","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Write the dataset card for product-search-v5","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.4,"slice":"core","lang":"en"}
|
||||
{"prompt":"Lay out an RFC for serving multiple embedding-model generations concurrently. Address version routing, index ownership, storage growth, mixed-version queries, cache isolation, tenant migrations, deletion propagation, quality comparisons, fallback, and a milestone sequence that avoids rebuilding every vector index at once.","purpose":"planning","secondary":null,"mixed":false,"difficulty":0.9,"slice":"core","lang":"en"}
|
||||
{"prompt":"Add streaming JSONL ingestion to the labeling CLI","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.5,"slice":"core","lang":"en"}
|
||||
{"prompt":"Create the annotation-review workspace with synchronized image zoom, hotkeys, issue tags, and reviewer progress.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Set TRAINING_SEED to 4821","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"core","lang":"en"}
|
||||
{"prompt":"Move normalization statistics loading behind one shared transform while preserving current floating-point outputs.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"Validation loss becomes NaN only with mixed precision and sequence lengths above 2048. Trace the first invalid operation.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Explain how the batch sampler balances languages and whether replacement is used. No edits.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.4,"slice":"core","lang":"en"}
|
||||
{"prompt":"Draft the model rollback runbook","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.3,"slice":"core","lang":"en"}
|
||||
{"prompt":"Propose a long-term architecture for privacy-preserving model evaluation on customer data. Include aggregation boundaries, minimum cohort sizes, deletion handling, reproducibility, experiment isolation, abuse prevention, auditability, regional constraints, failure modes, and phased validation before any results influence production promotion.","purpose":"planning","secondary":null,"mixed":false,"difficulty":1.0,"slice":"core","lang":"en"}
|
||||
{"prompt":"Implement checkpoint compaction for the distributed trainer so workers can resume without downloading every historical shard.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Add a responsive confusion-matrix viewer with normalized/raw modes, tooltips, zoom, and color-blind-safe palettes.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"Correct the dataset filename in configs/eval.yaml","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"core","lang":"en"}
|
||||
{"prompt":"Separate experiment assignment from metric emission without changing cohort membership or event fields.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"The feature backfill duplicates about 0.2% of rows after worker preemption. Determine which checkpoint or idempotency assumption fails.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Check whether the current vector-distance fallback is mathematically equivalent to the SIMD path for zero-length inputs.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"Escribe una guía para reproducir experimentos antiguos usando los snapshots de datos y el registro de modelos.","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.4,"slice":"core","lang":"es"}
|
||||
{"prompt":"Develop a roadmap for replacing the batch-only fraud model with hybrid streaming and batch features. Cover event-time handling, delayed labels, replay, online/offline consistency, cold starts, model refresh cadence, serving degradation, audit evidence, shadow evaluation, and independently reversible implementation milestones.","purpose":"planning","secondary":null,"mixed":false,"difficulty":0.9,"slice":"core","lang":"en"}
|
||||
{"prompt":"Build a Rust service that validates model artifacts, computes content hashes, and publishes immutable registry records.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Implement the feature-lineage graph with expandable nodes, search, impact highlighting, and keyboard navigation.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Change the chart label from 'AUC %' to 'AUC'","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Rename TrainingExample to LabeledExample across Python, Protobuf, and tests while preserving stored field names.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.7,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Why does the GPU worker silently switch to CPU after the third checkpoint? Training becomes unusably slow and the trigger is unknown.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.8,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Why is temperature scaling applied before tenant-specific calibration? Determine the intended current rationale; metrics are not known to be broken.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.5,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Document the existing prediction endpoint, including tensor formats, batching behavior, limits, and response examples.","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.5,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Plan the move from MLflow artifacts to an internal immutable registry, including data migration and rollout gates.","purpose":"planning","secondary":null,"mixed":false,"difficulty":0.8,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"The architecture in adr/042-online-features.md is approved; implement its Redis write-through path and version checks.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Add a compact model-status card with promotion state, health indicators, version history, and loading behavior.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.5,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Replace `trian_split` with `train_split` in one notebook utility","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Reorganize feature computation into source, transform, and sink layers while preserving rows, ordering, and metrics.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.7,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Investigate the unexplained ranking regression, restore correct behavior after locating the cause, and add a reproducible evaluation case.","purpose":"debugging","secondary":"backendImpl","mixed":true,"difficulty":0.9,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Review the current training-data deletion pipeline, then produce a compliance memo describing guarantees and gaps.","purpose":"review","secondary":"writing","mixed":true,"difficulty":0.8,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Write the inference API guide and add a runnable Python client that demonstrates streaming batches and retries.","purpose":"writing","secondary":"backendImpl","mixed":true,"difficulty":0.6,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Plan and implement the first stage of feature schema versioning: define compatibility rules, then add registry validation behind a warning-only mode.","purpose":"planning","secondary":"backendImpl","mixed":true,"difficulty":0.9,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Implement the drift-alert evaluator and document every emitted metric, threshold, and suppression rule.","purpose":"backendImpl","secondary":"writing","mixed":true,"difficulty":0.7,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Ticket MLUI-611\n\nResearchers compare up to six evaluation runs. The current page renders six separate full-width charts, forcing constant scrolling and making aligned time points impossible to inspect.\n\nRequirements:\n- metric selector supports accuracy, F1, latency, and calibration error\n- chosen runs share one x-axis but retain distinct series colors\n- hovering pins a vertical cursor across all visible plots\n- users can hide individual runs without losing selection\n- narrow screens use a horizontally scrollable plot, not compressed labels\n- exported PNG must include run names and dataset version\n\nImplement the comparison workspace in web/src/evaluation. Reuse our Vega wrapper and existing run query.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.8,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"Failing nightly command:\n\n$ python -m pipelines.evaluate --config configs/nightly.yaml\nFileNotFoundError: data/eval/golden-v6.parqet\n\nRepository paths:\ndata/eval/golden-v6.parquet\nconfigs/nightly.yaml: dataset: data/eval/golden-v6.parqet\nconfigs/release.yaml: dataset: data/eval/golden-v6.parquet\n\nThe nightly config contains a known misspelling. Correct only that path to `data/eval/golden-v6.parquet`. Do not regenerate data, edit the loader, or touch release.yaml.","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"Refactoring brief from maintainers:\n\ntrainer/checkpoint.py currently serializes model state, uploads blobs, updates registry metadata, prunes old checkpoints, and sends Slack alerts. The evaluator duplicates blob hashing and registry writes. Tests mock storage internals through CheckpointManager.\n\nDesired ownership:\n- artifact package: hashing and blob I/O\n- registry package: immutable metadata records\n- trainer: checkpoint policy\n- notifications: observes published events\n\nRestructure these modules and share artifact persistence with evaluator. Preserve checkpoint file bytes, registry schema, retention decisions, exception types, and CheckpointManager's public interface.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.8,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"Training log excerpt:\n\nstep=18820 loss=2.431 scale=8192 grad_norm=1.92\nstep=18821 loss=2.389 scale=8192 grad_norm=2.07\nstep=18822 loss=nan scale=4096 grad_norm=nan\nrank=3 attention_scores min=-inf max=inf seq_len=3072\nrank=0 attention_scores min=-31.2 max=28.7 seq_len=2048\nwarning: all_reduce completed with mismatched tensor shapes\n\nThe failure appears only when one rank receives a sequence longer than 2048 and gradient checkpointing is enabled. Find the first invalid shape or operation and explain why collective validation did not fail earlier. Diagnosis before code changes.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.9,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"Proposed sampling change:\n\ndiff --git a/data/sampler.py b/data/sampler.py\n@@\n- rng = random.Random(seed + epoch)\n- rng.shuffle(indices)\n+ rng = np.random.default_rng(seed)\n+ indices = rng.choice(indices, size=len(indices), replace=True)\n@@\n- return indices[rank::world_size]\n+ return np.array_split(indices, world_size)[rank]\n\nThe author says this only improves performance while preserving deterministic balanced sampling. Review determinism across epochs, replacement, shard sizes, duplicates, resume behavior, and compatibility with existing golden orders. Do not edit the patch.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.7,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"Write something for the model API","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.3,"slice":"vague-eval","lang":"en"}
|
||||
{"prompt":"Work out the platform direction","purpose":"planning","secondary":null,"mixed":false,"difficulty":0.6,"slice":"vague-eval","lang":"en"}
|
||||
{"prompt":"Finish the pipeline piece","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.5,"slice":"vague-eval","lang":"en"}
|
||||
{"prompt":"Build the Kubernetes rollout timeline with pod states, deployment events, revision markers, and auto-refresh pause controls.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Set the dev cluster region to eu-west-1","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"core","lang":"en"}
|
||||
{"prompt":"Extract provider-neutral DNS reconciliation from the Route53 controller without changing record behavior.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Pods remain Terminating for twenty minutes after a node disappears. Determine whether finalizers, volume detach, or eviction handling is stuck.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Assess the Terraform state bucket policy for cross-account access and deletion protection. Findings only.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Write the on-call guide for failed Argo rollouts","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.3,"slice":"core","lang":"en"}
|
||||
{"prompt":"Create a phased architecture plan for moving background jobs from Kubernetes CronJobs to a shared scheduler. Include tenancy, fairness, deadlines, retries, dependency graphs, secrets, regional failover, observability, migration adapters, rollback, and milestones that let individual teams opt in without coupling release schedules.","purpose":"planning","secondary":null,"mixed":false,"difficulty":0.9,"slice":"core","lang":"en"}
|
||||
{"prompt":"Implement a Go reconciler that renews external DNS leases with fencing tokens and conflict-safe retries.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Create the cloud-cost explorer with grouped filters, daily stacked charts, budget overlays, and drill-down tables.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Bump the base image to Debian 13","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"core","lang":"en"}
|
||||
{"prompt":"Deduplicate HTTP client construction across the AWS, GCP, and Azure adapters while preserving timeouts and retries.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"The ingress controller leaks file descriptors during certificate reloads and eventually stops accepting connections. Locate the unclosed resource.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Explain how preview environments receive isolated database roles and when those roles are revoked.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.5,"slice":"core","lang":"en"}
|
||||
{"prompt":"Rédige les notes de migration pour le nouveau chart Helm, avec les valeurs renommées et la procédure de rollback.","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.4,"slice":"core","lang":"fr"}
|
||||
{"prompt":"Architect a regional failover strategy for the deployment control plane. Address authoritative state, in-flight operations, fencing, artifact availability, DNS, credential scope, duplicate execution, operator control, degraded modes, recovery testing, and a rollout path from today's single-region system with measurable abort criteria.","purpose":"planning","secondary":null,"mixed":false,"difficulty":1.0,"slice":"core","lang":"en"}
|
||||
{"prompt":"Add per-namespace fairness to the Rust build queue with weighted limits and starvation metrics.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Implement a responsive policy-violation drawer with code snippets, remediation links, suppression controls, and keyboard focus restoration.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"Change replicas from 2 to 3","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"core","lang":"en"}
|
||||
{"prompt":"Move environment-name normalization into a shared package while keeping all accepted spellings and error strings unchanged.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.5,"slice":"core","lang":"en"}
|
||||
{"prompt":"Terraform reports perpetual diffs for IAM conditions even immediately after apply. Find the normalization or provider mismatch.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Walk through how the current admission webhook handles timeouts and failurePolicy settings across clusters.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.5,"slice":"core","lang":"en"}
|
||||
{"prompt":"Draft a runbook for rotating the container registry credentials","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.3,"slice":"core","lang":"en"}
|
||||
{"prompt":"Plan a migration from bespoke service dashboards to standardized OpenTelemetry service views. Include semantic conventions, legacy metric mapping, cardinality budgets, traces-to-logs correlation, ownership, SLO compatibility, dual-running, validation, training, and retirement milestones for old dashboards and exporters.","purpose":"planning","secondary":null,"mixed":false,"difficulty":0.9,"slice":"core","lang":"en"}
|
||||
{"prompt":"Build a controller that snapshots PersistentVolumeClaims on policy schedules and records restore metadata.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Add an environment-creation wizard with quota estimates, region compatibility warnings, progress states, and retryable failures.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Correct the Helm repository URL","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"core","lang":"en"}
|
||||
{"prompt":"Split policy evaluation, formatting, and CLI exit handling into separate modules with identical outputs and exit codes.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"A canary rollout occasionally promotes despite failed analysis runs. Trace the race between metric evaluation and deadline handling.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Review whether the current backup retention rules satisfy monthly and yearly restore requirements. No configuration changes.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"Create contributor docs for adding a new cloud provider adapter","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.4,"slice":"core","lang":"en"}
|
||||
{"prompt":"Define a roadmap for replacing long-lived CI cloud keys with federated identity. Cover trust configuration, repository scoping, pull requests from forks, reusable workflows, emergency fallback, auditability, phased enforcement, detection of remaining keys, and rollback if release pipelines cannot assume roles.","purpose":"planning","secondary":null,"mixed":false,"difficulty":0.9,"slice":"core","lang":"en"}
|
||||
{"prompt":"Implement an idempotent cleanup worker for expired preview stacks with dependency-aware deletion and resumable checkpoints.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Build the incident overview screen with service topology, active alerts, timeline annotations, and role-based action visibility.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Disable TRACE_EXPORT in local.env","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"core","lang":"en"}
|
||||
{"prompt":"Rename ClusterLease to ControlLease across services while preserving CRD fields and stored resource names.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"DNS records disappear briefly during no-op reconciliations in one region. Determine why the controller deletes before confirming replacements.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Examine the current network-policy generator and tell me whether empty peer lists mean deny-all or no restriction.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.5,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Document the current deployment API, including rollout phases, cancellation semantics, and retry examples.","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.5,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Plan the transition from static namespaces to virtual clusters; implementation belongs in later work.","purpose":"planning","secondary":null,"mixed":false,"difficulty":0.8,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Implement the approved backup manifest format and validation rules in backup-controller.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.6,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Erstelle die UI für den Rollback-Dialog mit Revisionsvergleich, Risikohinweis und klarer Bestätigung.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.5,"slice":"boundary","lang":"de"}
|
||||
{"prompt":"Set the warning icon to 16px","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Reorganize cloud credentials by provider without changing resolution order or supported environment variables.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.6,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Why does the scheduler launch two copies of a singleton job after leader failover? The duplicate execution is broken and unexplained.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.8,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Explain why the current autoscaler ignores pending pods with local volumes; nothing is currently reported as failing.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.5,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Write a plain-language overview of how infrastructure plans become applied changes, based on existing code.","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.4,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Design the migration to signed deployment manifests and implement signature verification in audit-only mode.","purpose":"planning","secondary":"backendImpl","mixed":true,"difficulty":0.9,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Add a capacity-report command and document its flags, JSON schema, and automation examples.","purpose":"backendImpl","secondary":"writing","mixed":true,"difficulty":0.6,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Create the secrets-rotation status page and add a support note explaining each displayed state.","purpose":"frontendImpl","secondary":"writing","mixed":true,"difficulty":0.6,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Fix the known staging account ID and update the assertion that copied the old value.","purpose":"quickFix","secondary":"backendImpl","mixed":true,"difficulty":0.3,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Consolidate resource label parsing with unchanged behavior, then document the accepted syntax for maintainers.","purpose":"refactor","secondary":"writing","mixed":true,"difficulty":0.6,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Infrastructure review excerpt:\n\nresource \"aws_iam_role\" \"preview_runner\" {\n assume_role_policy = data.aws_iam_policy_document.runner_assume.json\n}\n\ndata \"aws_iam_policy_document\" \"runner_assume\" {\n statement {\n actions = [\"sts:AssumeRole\"]\n principals { type = \"AWS\", identifiers = [\"*\"] }\n condition { test = \"StringLike\", variable = \"aws:PrincipalArn\", values = [\"arn:aws:iam::*:role/ci-*\"] }\n }\n}\n\nThe role policy grants CloudFormation, IAM PassRole, and S3 artifact access. Review whether the wildcard principal plus condition safely limits assumption, and assess confused-deputy, account-boundary, and PassRole risks. Findings only.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.8,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"Docs ticket OPS-1287\n\nThe restore command now has three phases:\n1. validate checks snapshot metadata and encryption-key access\n2. stage restores into an isolated namespace\n3. promote switches service routing after health checks\n\nFlags are `--snapshot`, `--target`, `--skip-promote`, and `--max-lag`. A failed promote leaves the staged database running for inspection. Running the same restore ID resumes from its last completed phase. Operators keep assuming retries start over and deleting the staged namespace.\n\nRewrite docs/runbooks/restore.md to explain phases, resume behavior, cleanup, flags, and abort conditions. No controller changes.","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.5,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"Planning notes from platform sync:\n\n- 430 repositories use static AWS keys in CI secrets.\n- 60 accept pull requests from forks.\n- reusable release workflows run in a central repository.\n- some self-hosted runners lack outbound access to the identity provider.\n- security wants keys disabled within six months.\n- release engineering requires a one-command emergency fallback.\n\nProduce a phased workload-identity migration plan. Define trust boundaries, repository enrollment, fork behavior, central workflow claims, offline-runner handling, telemetry, enforcement gates, emergency access, and rollback. Do not implement provider resources yet.","purpose":"planning","secondary":null,"mixed":false,"difficulty":0.9,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"Story PLATFORM-731\n\nCreate `platformctl snapshots prune`.\n\nRules:\n- retain the newest 14 daily, 8 weekly, and 12 monthly snapshots per database\n- snapshots under legal hold are never deleted\n- dry-run is default; `--execute` performs deletions\n- concurrent invocations must not delete the same snapshot twice\n- partial cloud failures should be resumable\n- JSON output includes kept/deleted/reason for every candidate\n- exit 2 when policy configuration is invalid, 1 for operational failure\n\nImplement the command and shared retention engine in Rust with golden tests and a fake object-store integration suite.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.8,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"Dashboard request:\n\nOperators need a live view of namespace quota pressure. Each row shows requested, used, and limit for CPU, memory, GPU, and persistent storage. Values update every 15 seconds. Above 80% use warning colors; above 95% use critical colors plus text, never color alone. Rows can expand to list the five largest workloads. Filters cover cluster, owner, and environment. On mobile, metrics stack inside cards. Implement this in console/src/quotas using the existing metrics endpoint and design tokens.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"Make the tiny UI adjustment","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.2,"slice":"vague-eval","lang":"en"}
|
||||
{"prompt":"Fix the config bit","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"vague-eval","lang":"en"}
|
||||
{"prompt":"Split the CLI argument parser from command execution while keeping flags, help output, and exit codes unchanged.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"The Rust formatter enters an infinite loop on nested comments followed by a raw string. Find the parser state that stops advancing.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Review the bytecode verifier for paths that could read beyond a function body. No modifications.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Write shell completion setup docs","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.2,"slice":"core","lang":"en"}
|
||||
{"prompt":"Design a migration from the current AST interpreter to a register-based bytecode VM. Cover instruction format, source maps, exceptions, garbage collection boundaries, native extensions, debugging, compatibility tests, incremental adoption, performance gates, and rollback while the interpreter remains available.","purpose":"planning","secondary":null,"mixed":false,"difficulty":1.0,"slice":"core","lang":"en"}
|
||||
{"prompt":"Implement response-file expansion for the C++ compiler driver with cycle detection and platform-specific quoting.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Create a terminal dependency-tree viewer with folding, search, color themes, and screen-reader-friendly plain output.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"Change the default output extension to .wasm","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"core","lang":"en"}
|
||||
{"prompt":"Consolidate duplicate span-merging logic across lexer and macro expander without changing diagnostic locations.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Incremental builds sometimes reuse stale generated headers after a feature flag changes. Determine which dependency edge is missing.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Explain how the linker chooses between weak symbols from two archives in the current implementation.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"Escreva a referência dos códigos de saída do comando `forge test`.","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.3,"slice":"core","lang":"pt"}
|
||||
{"prompt":"Plan a plugin architecture for the static analyzer that allows third-party rules without exposing unstable compiler internals. Address process isolation, version negotiation, diagnostics, configuration, caching, security, distribution, failure handling, and milestones for migrating built-in rules before opening the interface externally.","purpose":"planning","secondary":null,"mixed":false,"difficulty":0.9,"slice":"core","lang":"en"}
|
||||
{"prompt":"Add a streaming parser for large SARIF reports with bounded memory and preserved result ordering.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Build the TUI test runner with live worker status, failure expansion, rerun shortcuts, and monochrome support.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Fix `--quite` to `--quiet` in one example","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"core","lang":"en"}
|
||||
{"prompt":"Extract target-triple normalization into a shared crate while preserving all aliases and diagnostics.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"The compiler crashes only when an async closure captures a generic associated type. Reduce and locate the invalid lowering assumption.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.9,"slice":"core","lang":"en"}
|
||||
{"prompt":"Check whether the package resolver's tie-breaking is deterministic when two registries contain the same version.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"Prepare a contributor guide for adding a lint rule","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.3,"slice":"core","lang":"en"}
|
||||
{"prompt":"Outline an architecture for reproducible remote builds across Linux, macOS, and Windows. Include toolchain identity, hermetic inputs, cache keys, sandboxing, secrets, network access, artifact provenance, platform-specific signing, failure fallback, rollout phases, and objective comparisons against current local builds.","purpose":"planning","secondary":null,"mixed":false,"difficulty":1.0,"slice":"core","lang":"en"}
|
||||
{"prompt":"Implement content-addressed artifact caching in the Go build daemon with concurrent upload deduplication.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Add a web-based flamegraph viewer with zoom, search, frame details, and shareable selections.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"Bump the language edition to 2024","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"core","lang":"en"}
|
||||
{"prompt":"Move package lockfile encoding behind a versioned serializer without changing current bytes.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"A cancellation signal causes some test workers to hang while others exit. Find the descriptor or task preventing shutdown.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Walk through how macro hygiene identifiers are generated and compared in the current compiler.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"Draft release notes for forge 3.0","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.3,"slice":"core","lang":"en"}
|
||||
{"prompt":"Create a migration roadmap for replacing bespoke configuration parsing with a typed schema system. Address backward compatibility, unknown keys, environment interpolation, secret references, generated docs, editor support, plugin fields, automated conversion, warning phases, and rollback before legacy syntax is rejected.","purpose":"planning","secondary":null,"mixed":false,"difficulty":0.9,"slice":"core","lang":"en"}
|
||||
{"prompt":"Implement a daemon protocol for multiplexing compiler requests with cancellation and per-client resource quotas.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Build the package graph screen with cycle highlighting, version badges, edge filtering, and SVG export.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Set the formatter line width to 100","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"core","lang":"en"}
|
||||
{"prompt":"Rename BuildUnit to CompilationUnit across the codebase while preserving cache record fields and plugin APIs.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"The dependency solver picks an older incompatible package only when optional features form a cycle. Diagnose the search-state bug.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Review whether the existing response-file parser matches Windows CommandLineToArgvW semantics. Do not change it.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.7,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Document the current compiler JSON diagnostics format with examples, not a redesign.","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.4,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Plan the migration to a new lockfile format, including staged compatibility; implementation comes later.","purpose":"planning","secondary":null,"mixed":false,"difficulty":0.8,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Implement the finalized module-resolution algorithm from specs/modules-v2.md.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"为命令行性能报告做一个网页界面，要有火焰图、筛选器、对比模式和可分享链接。","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.6,"slice":"boundary","lang":"zh"}
|
||||
{"prompt":"Replace `--colour` with `--color` in docs/quickstart.md","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Rearrange parser passes into explicit stages without changing accepted programs or diagnostics.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.7,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Why does `forge test` report success when a worker is killed by OOM? The result aggregation is broken and the cause is unknown.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.7,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Explain why the current linker keeps discarded section names in the string table; no output bug is reported.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.5,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Write an overview of the existing incremental compilation cache for plugin authors.","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.5,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Plan the bytecode verifier rollout and implement a warning-only verifier pass for unsupported instructions.","purpose":"planning","secondary":"backendImpl","mixed":true,"difficulty":0.9,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Add a `forge explain` command and document its query syntax, output schema, and examples.","purpose":"backendImpl","secondary":"writing","mixed":true,"difficulty":0.7,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Create the browser-based diagnostic explorer and write a QA checklist for large files and keyboard navigation.","purpose":"frontendImpl","secondary":"writing","mixed":true,"difficulty":0.7,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Correct the known default target triple and update the unit test asserting the old value.","purpose":"quickFix","secondary":"backendImpl","mixed":true,"difficulty":0.3,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Consolidate package-name validation without behavior changes, then document the shared invariants.","purpose":"refactor","secondary":"writing","mixed":true,"difficulty":0.6,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Code review excerpt:\n\nfn verify_jump(func: &Function, pc: usize, offset: i32) -> Result<()> {\n let target = pc.wrapping_add(offset as usize);\n if target > func.code.len() {\n return Err(Error::InvalidJump(target));\n }\n if func.code[target].is_operand_byte() {\n return Err(Error::JumpIntoOperand(target));\n }\n Ok(())\n}\n\nBytecode offsets are signed relative to the next instruction. Empty functions are allowed in malformed inputs and must be rejected safely. Review integer conversion, boundary checks, indexing, backward jumps, and the stated offset base. Do not patch it; provide exploitable and correctness findings.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.8,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"Documentation issue CLI-441\n\nUsers confuse these commands:\n- `forge cache clean` removes local entries only\n- `forge cache purge --remote` requests remote deletion\n- `forge cache gc` removes unreachable local blobs after scanning lockfiles\n\nRemote purge is asynchronous and returns request ID. GC may retain blobs referenced by worktrees outside the current directory if they are registered. Clean never affects build outputs already copied into dist/.\n\nRewrite docs/cache.md with a comparison table, safety notes, examples, and a section explaining what each command does not delete. No behavior changes.","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.4,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"VM architecture notes:\n\n- Interpreter executes roughly 65 million instructions/sec.\n- Startup must remain under 20 ms for CLI scripts.\n- Native extensions can hold opaque interpreter object pointers.\n- Debugger breakpoints refer to AST node IDs.\n- Existing snapshot tests cover 18,000 programs.\n- We need bytecode serialization eventually, but not in phase one.\n\nTurn this into a staged design for a register-based VM. Define instruction representation, lowering, exception semantics, GC interaction, native-extension compatibility, debugger mapping, performance gates, validation, and rollback to the interpreter.","purpose":"planning","secondary":null,"mixed":false,"difficulty":1.0,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"Issue FORGE-1204\n\nSupport response files prefixed with `@`.\n\nRequirements:\n- UTF-8 with optional BOM on Unix\n- Windows decoding follows the active code page unless BOM is present\n- nested files resolve relative to the containing file\n- maximum nesting depth 16\n- report the full inclusion chain on cycles\n- quoting follows the platform command-line parser\n- `@@foo` escapes to literal `@foo`\n- expansion occurs before subcommand parsing\n\nImplement this in compiler-driver and add cross-platform golden tests. Existing direct argument parsing must remain unchanged.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.8,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"Web tooling ticket:\n\nThe CI trace viewer currently renders 50,000 spans as DOM rows and freezes Chrome. Replace the main pane with a virtualized timeline. Users need zoom, pan, text search, critical-path highlighting, worker-lane collapse, duration tooltips, and a details drawer. Selection must survive live trace updates. Provide a plain table fallback when reduced motion or assistive technology mode is enabled. The API already returns normalized spans; don't change it.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.8,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"Change the little option","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"vague-eval","lang":"en"}
|
||||
{"prompt":"Clean up compiler stuff","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.5,"slice":"vague-eval","lang":"en"}
|
||||
{"prompt":"Figure out why builds are weird","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.6,"slice":"vague-eval","lang":"en"}
|
||||
{"prompt":"Audit the payment webhook handler for replay, ordering, and signature-validation risks. Findings only.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Write the PCI scope overview for engineers","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.4,"slice":"core","lang":"en"}
|
||||
{"prompt":"Design a migration from synchronous card authorization to an asynchronous payment-intent workflow. Cover state ownership, idempotency, webhook races, client compatibility, partial failures, ledger timing, fraud checks, observability, reconciliation, rollout milestones, and rollback without double-charging customers.","purpose":"planning","secondary":null,"mixed":false,"difficulty":1.0,"slice":"core","lang":"en"}
|
||||
{"prompt":"Implement a Java service that reserves gift-card balances atomically across split-tender checkout.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Create the checkout payment-method picker with saved cards, wallets, gift cards, validation, and accessible errors.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Set the test currency to EUR","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"core","lang":"en"}
|
||||
{"prompt":"Extract money rounding into a shared value type while preserving every historical invoice result.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Some refunds settle twice after a webhook retry, despite an idempotency key. Locate the transaction or key-scope defect.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.9,"slice":"core","lang":"en"}
|
||||
{"prompt":"Check whether the current payout scheduler can starve low-volume merchants. Do not change scheduling.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"Schreibe eine Integrationsanleitung für signierte Zahlungs-Webhooks mit Fehlerbeispielen.","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.4,"slice":"core","lang":"de"}
|
||||
{"prompt":"Map out an architecture for marketplace seller payouts across currencies and regions. Include ledger invariants, reserves, negative balances, FX timing, tax holds, banking cutoffs, retries, reconciliation, regulatory boundaries, operational overrides, and staged delivery starting with one country and one payout rail.","purpose":"planning","secondary":null,"mixed":false,"difficulty":1.0,"slice":"core","lang":"en"}
|
||||
{"prompt":"Add durable idempotency storage to the TypeScript payment-intent service with payload-conflict detection.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Build the merchant dispute dashboard with evidence status, deadlines, amount filters, and bulk document upload.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Correct the cents conversion constant","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"core","lang":"en"}
|
||||
{"prompt":"Consolidate duplicated webhook event parsing across Stripe and Adyen adapters without changing normalized events.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"The reconciliation job shows balances off by one cent only for JPY refunds converted through USD. Find the rounding sequence error.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Explain how the current ledger represents authorization holds, captures, reversals, and expirations.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.5,"slice":"core","lang":"en"}
|
||||
{"prompt":"Draft customer-facing release notes for bank-transfer payments","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.3,"slice":"core","lang":"en"}
|
||||
{"prompt":"Plan the retirement of a legacy payment processor while preserving subscriptions, stored payment references, refunds, disputes, reporting, and historical reconciliation. Include processor coexistence, merchant cohorting, customer communication, data portability, shadow comparisons, rollback, and criteria for shutting down old webhooks.","purpose":"planning","secondary":null,"mixed":false,"difficulty":0.9,"slice":"core","lang":"en"}
|
||||
{"prompt":"Implement the ACH return-event consumer with deduplication and ledger compensation entries.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Add an invoice-payment timeline with state icons, expandable events, timezone display, and retry feedback.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"Bump the sandbox API version","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"core","lang":"en"}
|
||||
{"prompt":"Move processor-specific decline mapping behind an adapter while retaining all customer-visible messages.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Payment sessions occasionally remain pending after both capture and webhook success. Determine which state transition is lost.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Review whether the current card-fingerprint handling can link customers across tenants unintentionally.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Write a support runbook for delayed payouts","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.3,"slice":"core","lang":"en"}
|
||||
{"prompt":"Propose a design for real-time transaction risk scoring that remains available when the ML service is degraded. Cover latency budgets, feature freshness, deterministic fallbacks, manual rules, explainability, audit retention, shadow evaluation, regional routing, rollout, and rollback without changing approval rates abruptly.","purpose":"planning","secondary":null,"mixed":false,"difficulty":0.9,"slice":"core","lang":"en"}
|
||||
{"prompt":"Build a reconciliation command that compares processor settlements to ledger entries and emits resumable discrepancy files.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Implement the mobile refund flow with amount selection, reason entry, confirmation, progress, and partial-refund history.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Change PAYMENT_TIMEOUT_MS to 45000","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"core","lang":"en"}
|
||||
{"prompt":"Rename SettlementBatch to PayoutBatch across services while preserving database tables and event names.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Card verification fails only when customers return from a specific bank's 3DS page. Diagnose the malformed or lost redirect state.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.8,"slice":"core","lang":"en"}
|
||||
{"prompt":"Read the chargeback evidence builder and tell me which fields are omitted for digital goods.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.5,"slice":"core","lang":"en"}
|
||||
{"prompt":"Prepare an API reference for payout status events","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.4,"slice":"core","lang":"en"}
|
||||
{"prompt":"Create a migration plan for adopting ISO 20022 bank messages, covering schema mapping, coexistence, reconciliation, validation, partner certification, observability, archival requirements, phased corridor rollout, and rollback to legacy formats.","purpose":"planning","secondary":null,"mixed":false,"difficulty":0.9,"slice":"core","lang":"en"}
|
||||
{"prompt":"Add a fraud-rule evaluator with versioned rulesets and deterministic decision traces.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Build the payment-link editor with live preview, expiration controls, branding, and localization-safe layout.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.6,"slice":"core","lang":"en"}
|
||||
{"prompt":"Replace the outdated support email","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"core","lang":"en"}
|
||||
{"prompt":"Separate invoice rendering from tax calculation while preserving PDFs, totals, and event timing.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Payout reports intermittently omit the final transaction of each page. Find whether cursor construction or ordering is wrong.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.7,"slice":"core","lang":"en"}
|
||||
{"prompt":"Assess whether the existing retry policy can duplicate non-idempotent processor calls. Analysis only.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.7,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Document the current refund API behavior, including partial refunds and idempotent retries.","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.5,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Plan the migration from floating-point tax calculations to decimal arithmetic; do not implement it yet.","purpose":"planning","secondary":null,"mixed":false,"difficulty":0.8,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Implement the approved payout state machine from docs/payout-states.md.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Criar a tela de confirmação de pagamento com resumo, taxas, cupom, erros e suporte a leitor de tela.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.6,"slice":"boundary","lang":"pt"}
|
||||
{"prompt":"Change 'succesful' to 'successful' in Receipt.tsx","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Restructure the ledger posting code into validation, entry construction, and persistence with identical results.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.7,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Why do some expired authorization holds remain in available-balance calculations? The balance is wrong and the source is unknown.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.8,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Explain why the current webhook verifier accepts two active signing secrets during rotation. Nothing is failing.","purpose":"review","secondary":null,"mixed":false,"difficulty":0.4,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Write an operational overview of existing settlement reconciliation, not a proposal.","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.5,"slice":"boundary","lang":"en"}
|
||||
{"prompt":"Plan the ledger partitioning migration and implement a read-only analyzer that estimates tenant movement cost.","purpose":"planning","secondary":"backendImpl","mixed":true,"difficulty":0.9,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Implement the dispute-evidence endpoint and document its file limits, states, and error responses.","purpose":"backendImpl","secondary":"writing","mixed":true,"difficulty":0.7,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Build the subscription pause UI and add a product QA note covering dates, proration copy, and accessibility.","purpose":"frontendImpl","secondary":"writing","mixed":true,"difficulty":0.6,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Fix the known tax-region code and update the backend fixture that contains the old code.","purpose":"quickFix","secondary":"backendImpl","mixed":true,"difficulty":0.3,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Consolidate processor error translation without behavior changes, then document the mapping contract.","purpose":"refactor","secondary":"writing","mixed":true,"difficulty":0.7,"slice":"mixed","lang":"en"}
|
||||
{"prompt":"Payment implementation ticket:\n\nPOST /v2/payment-intents/{id}/capture\nBody: { amount?, idempotency_key }\n\nRules:\n- omitted amount captures the remaining authorized balance\n- partial capture is allowed up to three times\n- total captured cannot exceed authorization\n- declined captures do not consume the idempotency key\n- successful responses replay exactly for 48 hours\n- processor call and ledger posting cannot share a transaction\n- publish CaptureCompleted only after ledger durability\n- concurrent captures must serialize per intent\n\nImplement the command handler, locking, processor adapter call, recovery state, outbox event, and integration tests for concurrency and retries.","purpose":"backendImpl","secondary":null,"mixed":false,"difficulty":0.9,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"Checkout design ticket:\n\nAt review step, users need to see subtotal, discounts, tax, shipping, gift-card credit, and final amount. If the total changes after address validation, highlight changed rows and require acknowledgment before Pay becomes enabled. On phones the summary is collapsible but the final amount stays visible. Errors from wallets should appear beside the selected method, not as a global toast. Support 200% zoom, screen readers, and right-to-left currencies. Implement in CheckoutReview.tsx with existing pricing data; no backend changes.","purpose":"frontendImpl","secondary":null,"mixed":false,"difficulty":0.7,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"CI snapshot failure:\n\n- Expected\n+ Received\n\n- \"[email protected]\"\n+ \"[email protected]\"\n\nProduct copy file:\nconfig/contact.json currently contains `[email protected]`.\nApproved ticket value: `[email protected]`.\nThe snapshot was already updated in the branch, but the source value was missed. Change only config/contact.json to the approved address. Do not modify the snapshot or tests.","purpose":"quickFix","secondary":null,"mixed":false,"difficulty":0.1,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"Refactoring request from ledger team:\n\nposting/service.ts validates commands, converts currencies, builds entries, opens transactions, persists entries, updates cached balances, and publishes events. The reconciliation worker duplicates entry construction for imported settlements. Tests rely on exact entry order and error types.\n\nCreate explicit validation, entry-builder, and persistence boundaries; share the builder with reconciliation. Preserve transaction scope, journal ordering, rounding, public service methods, event payloads, balance timing, and every existing error code. This is structural work only—no new ledger states or processor behavior.","purpose":"refactor","secondary":null,"mixed":false,"difficulty":0.8,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"Incident log:\n\n09:18:41 refund request id=r88 amount=4000 idempotency=abc\n09:18:41 processor accepted external=rf_991\n09:18:42 db commit timeout connection reset\n09:18:44 retry id=r88 idempotency=abc lookup miss\n09:18:44 processor accepted external=rf_992\n09:18:45 ledger posted refund=rf_992\n09:19:02 webhook received external=rf_991\n\nTwo processor refunds occurred, but our idempotency table contains only rf_992. Determine the failure window and why retry lookup missed the first attempt. Reproduce it and identify the correct recovery invariant before changing code.","purpose":"debugging","secondary":null,"mixed":false,"difficulty":0.9,"slice":"pasted-context","lang":"en"}
|
||||
{"prompt":"Tell me if payments look okay","purpose":"review","secondary":null,"mixed":false,"difficulty":0.4,"slice":"vague-eval","lang":"en"}
|
||||
{"prompt":"Write the usual blurb","purpose":"writing","secondary":null,"mixed":false,"difficulty":0.2,"slice":"vague-eval","lang":"en"}
|
||||
Reference in New Issue
Block a user