Coding standards, frameworks, refactoring, and code review (18.1K skills)
Review an XQUIC agent harness for duplicated guidance, overlapping skill triggers, broken references, and missing automated checks. Get prioritized findings, policy decisions, proposed document or script changes, and a record of validation performed or still needed.
Check whether a proposed XQUIC bug is supported by an applicable RFC or Internet-Draft and the current source behavior. The skill traces evidence, compares expected and actual logic, and returns a fail-closed decision that can gate issue submission.
Prepare a focused XQUIC bug report with reproducible steps, comparable expected and actual behavior, and an exact RFC or draft reference. It submits the report only after the separate issue-check workflow verifies the claim and other submission gates are clear.
Assess how much work an XQUIC issue may take to resolve and assign an evidence-based L1–L4 complexity level. The workflow records validity, priority, severity, and disposition separately, and applies labels only with high confidence and explicit authorization.
Stage, commit, and push XQUIC changes with scoped Git checks. Review staged and unstaged files, confirm the branch and resolved remote target, preserve unrelated work, and follow repository commit conventions before publishing a branch.
Prepare and update concise pull request titles and bodies for alibaba/xquic, following its contribution format. Check that the summary matches the current code and validation evidence, create draft PRs from pushed branches, and verify state before moving a complete PR toward review.
Prepare and run a local XQUIC build on macOS or Linux. The skill checks the platform and build options, then uses the repository validation script to compile targets and report the build setup and any failures.
Review one or more pull requests or code changes against criteria for security, correctness, clean code, TypeScript, React, localization and right-to-left layouts, performance, accessibility, and testing. It organizes findings by severity and category, including prior feedback on re-reviews.
Apply a prioritized set of React and Next.js performance guidelines when building, reviewing, or refactoring components and pages. The rules cover data fetching, bundle size, server and client work, re-renders, rendering, and JavaScript, with examples of better patterns.
Prepare a new skill for release by following a checklist for its documentation, logo, metadata, and project listings. Use it when adding or publishing a skill, or when checking that required release materials are in place.
Use this template to build a new agent skill. It outlines the required frontmatter, documentation sections, optional scripts and examples, and the steps for adding the skill to a project. It provides structure rather than a specific skill capability.
Use this skill when changing the open-websearch MCP server. It helps classify engine, networking, or tool changes, keep fixes within the affected area, preserve tool contracts, and check relevant tests and documentation before completing the work.
Plan and validate PostgreSQL schema migrations for live or shared databases. Learn which DDL operations lock tables, how to batch backfills, create indexes concurrently, test against a database fork, set timeouts, and prepare rollback steps.
Build PostgreSQL search that combines BM25 keyword matches with vector-based semantic results. The skill covers parallel queries, Reciprocal Rank Fusion (a way to merge ranked lists), weighting, and optional machine-learning reranking for candidate results.
Use this reference when creating or changing PostgreSQL tables and schemas. It explains how to choose data types, constraints, and indexes, and covers JSONB, partitioning, and patterns for update-heavy or upsert-heavy workloads.
Use this skill for PostgreSQL work such as inspecting schemas, designing tables and indexes, building vector or spatial search, and planning migrations. It points to task-specific guidance and SQL patterns for each area.
Design PostGIS tables for location-based applications by choosing spatial types, coordinate systems, constraints, and indexes. Examples cover points, routes, parcels, and service areas, with query and validation patterns for common spatial workloads.
Use this reference to reason about how concurrent database transactions interact. It helps classify conflicts and isolation anomalies, compare protocol families such as locking, MVCC, and OCC, and assess whether a commit or abort is required for correctness.
Use `tshark` to inspect PCAP captures, filter HTTP traffic, extract request fields, follow TCP conversations, and export payload bytes. The workflow helps narrow a capture from broad protocol searches to specific flows and details for analysis.
Organize a Three.js scene into named parts, apply world transforms to mesh geometry, and export each part as an OBJ file. The workflow can also build a URDF model from named links, with deterministic ordering for repeatable output.
Plan and build a Codex skill for a specialized task, including its instructions and optional scripts, references, or assets. The guide covers initialization, editing, packaging, and iteration so the result is organized for reuse.
Plan a migration from Spring Boot 2.x to 3.x with guidance on version and Java updates, incompatible dependencies, and javax-to-jakarta changes. The skill includes an OpenRewrite migration example and a checklist for compiling and reviewing the application afterward.
Update Spring Security 5 configurations for Spring Security 6 by replacing the adapter-based setup with SecurityFilterChain beans, changing URL matcher and method-security APIs, and converting chained configuration to the lambda DSL. It also highlights authentication manager and servlet namespace changes.
Follow the build workflow for adding syscall descriptions to syzkaller: edit description files, provide constants, compile descriptions, and build the project. It also outlines quick rebuilds, common errors, and places to find existing examples.