개발 스킬

Review agent instructions, skills, tool definitions, and linked docs in a repository. Find conflicts, outdated facts, repeated guidance, and missing context, then produce a scored report with concrete rewrite suggestions.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Reviewing a repository’s CLAUDE.md and agent docs
• Finding conflicting or outdated instructions
• Reducing repeated context across skills and tools

Prepare a repository for agent-driven development by improving its documentation map, local development setup, end-to-end checks, verification workflow, and custom lints. The skill assesses what is missing and builds capabilities incrementally, with commit hygiene and ongoing cleanup practices.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Onboarding an unfamiliar codebase for agent work
• Setting up a reproducible local development stack
• Adding end-to-end verification and repository-specific lints

Port C, C++, Python, Go, JavaScript, or TypeScript code to idiomatic Rust while preserving observable behavior. Map types and errors to Rust patterns, carry over or create tests, verify the result when tools are available, and document API mappings and deliberate differences.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Rewrite a Python module in Rust while preserving behavior
• Move C error codes and memory handling into Rust types
• Port existing tests alongside a Rust implementation

Use this skill to write, review, refactor, debug, or explain Rust code. It applies guidance on ownership, types, errors, APIs, async, unsafe code, and project structure to help produce idiomatic, robust code that is easier to maintain.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Starting a Rust crate or module
• Reviewing or refactoring Rust code
• Understanding ownership or lifetime issues

Review Rust files, modules, crates, or diffs for correctness, safety, error handling, API design, idioms, and maintainability. Get prioritized findings with cited rules, locations, concrete code fixes, and a summary of what is already working well.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Review a Rust module before merging a change
• Check a crate's public API and error handling
• Find safety issues and idiom improvements in a diff

Review DataFusion Comet pull requests that change native or JVM shuffle. Check partitioning behavior, block formats, memory and spilling, writer selection, Celeborn paths, tests, and contributor documentation for compatibility risks.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Reviewing a change to native shuffle partitioning
• Checking a JVM shuffle writer or spill change
• Reviewing Arrow IPC block format or compression edits

Review DataFusion Comet pull requests that change memory handling. Check which memory budget is affected, how reservations and pools behave, and how Spark and native code interact, then identify missing evidence or documentation updates.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Reviewing a Comet PR that changes memory reservations
• Checking a change to spilling or memory pools
• Assessing Spark TaskMemoryManager bridge changes

Review DataFusion Comet changes that cross the JVM/native boundary, focusing on Arrow data transfer, ownership, lifetimes, offsets, and memory release. It helps identify risks such as leaks, invalid access, or crashes and is used alongside the general Comet PR review skill.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Review a PR changing Arrow C Data or C Stream handling
• Check batch ownership and cleanup across JVM and native code
• Inspect vector export, import, offsets, and lifecycle changes

Connect an existing DataFusion or datafusion-spark function to a Spark expression in Comet. Choose a direct mapping, UDF registration, or custom serde for input handling and restrictions, then add SQL file tests. It assumes the upstream function already exists.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Connecting an existing function to a Spark expression
• Registering a datafusion-spark UDF for Comet
• Adding custom input handling or type restrictions

Assess whether a Spark expression routed through Comet’s JVM codegen dispatcher merits a native Rust path. Weigh exact Spark compatibility against likely throughput or allocation gains, record the verdict in the expression audit log, and file a GitHub issue for the recommendation.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Choosing the next expression for native implementation
• Evaluating a specific codegen-dispatched expression
• Recording why an expression was disqualified

Review a DataFusion Comet pull request with consistent checks for compatibility, release behavior, tests, CI, and documentation. The skill identifies which area-specific review guides apply and prepares findings and suggested comments for a human reviewer.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Reviewing a Comet pull request across multiple subsystems
• Checking a change against Spark behavior and the latest release
• Preparing actionable findings without posting a review

Review DataFusion Comet changes to Iceberg writes, including the split write plan, native writer, eligibility checks, and iceberg-rust updates. Assess output parity with iceberg-java, fallback behavior, cleanup on failure, test evidence, and documentation needs.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Reviewing changes to the native Iceberg writer
• Checking which Iceberg writes qualify for Comet
• Assessing cleanup and failure handling in write changes

Review DataFusion Comet pull requests that add or change Spark expressions. Check the Scala serialization path, Rust implementation, return types, edge cases, tests, and performance against Spark behavior, using the general Comet PR review guide for shared review steps.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Reviewing a new Spark expression implementation
• Checking expression behavior against Spark source and tests
• Assessing expression edge cases, test coverage, and benchmarks

Improve an existing native scalar expression in DataFusion Comet by benchmarking varied input shapes, preserving identical outputs, and checking for regressions. It guides changes to per-row or per-batch expression work such as casts, strings, JSON, arrays, and math.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Benchmark an existing scalar expression before and after a change
• Optimize string, JSON, array, cast, or math kernels
• Check correctness and regression risk across input shapes

Implement a Spark expression in DataFusion Comet by studying Spark’s behavior, reusing an upstream datafusion-spark function when possible, and wiring the Scala serde and Rust function. Add SQL tests, then use an audit workflow to find and address coverage gaps.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Adding a new Spark expression to Comet
• Checking whether an upstream function can be reused
• Adding serde and Rust expression wiring

Classify open Comet issues awaiting triage as bugs or enhancements, apply the project's type, priority, and area labels, and record decisions and skipped issues in a dated summary issue for human review.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Sorting new Comet issues into bugs and enhancements
• Applying priority and area labels to issues
• Preparing a triage summary for human review

Construct a deterministic `tanstack create` command for a new app by selecting its framework, build tool, template, toolchain, deployment target, and add-ons. The skill accounts for incompatible options and non-interactive defaults so the generated scaffold better matches the requested setup.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Generating a TanStack app without interactive prompts
• Choosing build tools and deployment targets for a scaffold
• Adding specific integrations to a generated app

Use TanStack CLI to discover valid library identifiers, fetch versioned documentation, search docs, and inspect available add-ons in machine-readable form. This gives an agent authoritative project context and helps validate library or integration choices before generating code.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Finding valid TanStack library identifiers
• Fetching a specific documentation page or version
• Searching docs and inspecting available CLI add-ons

Use this skill to choose valid TanStack CLI integrations for a new app. It maps ecosystem partners to installable add-on IDs, checks provider options, and accounts for exclusive categories and router-only compatibility limits.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Mapping an ecosystem partner to an installable add-on
• Checking provider options before choosing a data layer
• Avoiding conflicting integrations in a new app

Add integrations to an existing TanStack CLI project. Resolve valid add-on IDs, check required dependencies and options, and account for the project metadata the command needs so the additions can be configured correctly.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Adding Clerk or Drizzle to an existing app
• Checking available add-on IDs before adding one
• Reviewing an add-on’s options before installation

Use this skill to set up and iterate on custom TanStack add-ons. It guides initialization, compilation, development loops, and syncing watched framework files into a sandbox app, while checking project metadata and valid watch paths.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• カスタムアドオンを初期化したいとき
• 開発中の変更を検証用アプリへ同期したいとき
• アドオン開発の前提条件を確認したいとき

Combine the commits for one change into a single commit, run the project’s Gerrit review checks, and push the result to Gerrit drafts. The workflow also handles staged or unstaged tracked changes and preserves Gerrit commit-message trailers.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Prepare a multi-commit change for Gerrit review
• Squash commits before pushing a change
• Run the Gerrit critique check before submitting

Use this skill to compile one C++ source file in the Impala repository as a quick error check. It normalizes the filename, verifies the build command with a dry run, and reports the build result. It requires the configured Impala environment and its provided script.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Checking one C++ file for compile errors
• Validating a source basename before building
• Running a focused build without compiling the full project

Use this skill to plan and adapt HDInsight Spark, Hive, and Oozie workloads for Microsoft Fabric. It maps legacy APIs and storage paths to Fabric equivalents, converts Hive tables to Delta Lake, and outlines how to replace Oozie workflows with data pipelines.

한국어 요약은 준비 중입니다. 영어 요약을 표시하고 있습니다.

• Migrating HDInsight Spark notebooks to Fabric
• Replacing Hive tables with Lakehouse Delta tables
• Moving WASB or ABFS paths to OneLake shortcuts