Total 50,405 skills, Code Quality has 2287 skills
Showing 12 of 2287 skills
Profile-driven performance optimization with behavior proofs. Use when: optimize, slow, bottleneck, hotspot, profile, p95, latency, throughput, or algorithmic improvements.
Conduct an architecture health check on a design — either verify if the design is internally consistent (no conflicts between terminology, contracts, and implementation steps) or check if the design aligns with the code (ensuring what was promised in the design is actually implemented in code). This skill only outputs issue lists and repair suggestions, and does not make any modifications. It focuses on only one target each time; "顺手把另一项也查了" is not allowed. Trigger scenarios: Users say "perform architecture check", "is the design internally consistent?", "does the plan align with the code?", or want to conduct a health check before proceeding to the implement/acceptance phase.
MUST be used when user asks to document TypeScript or Vue code — add TSDoc/JSDoc blocks, docblocks, или «задокументировать» functions, classes, methods, interfaces, type aliases, enums, generics, Vue composables, defineProps/defineEmits/defineModel. Enforces strict TSDoc spec (tsdoc.org) and writes all descriptions in Russian. Triggers on phrases like «добавь tsdoc», «задокументируй», «напиши док», «add tsdoc», «document this», «generate jsdoc».
Stage 2 code quality review. Triggers: 'quality review', 'check code quality', or /review stage 2. Requires spec-review to have passed first. Checks SOLID, DRY, security, and test quality. Do NOT use for spec compliance — use spec-review instead.
Identify codebase deepening opportunities based on the domain language in CONTEXT.md and decisions in docs/adr/. Use when the user wants to improve architecture, find refactoring opportunities, consolidate tightly-coupled modules, or make a codebase more testable and AI-navigable.
Four-mantra debugging discipline — reproduce, trace the fail path, falsify the hypothesis, cross-reference every breadcrumb. Recite the mantra block verbatim at the start of any debugging session, then apply the four steps in order before proposing any fix. Trigger on /debug-mantra and proactively whenever debugging starts — user reports a bug, says something is broken/throwing/failing, asks to debug/diagnose/investigate an issue, or pastes a stack trace or error log.
Use this whenever you encounter any bugs, test failures, or abnormal behavior, and execute it before proposing a fix
Use after receiving code review feedback and before implementing suggestions, especially when feedback is unclear or technically questionable—requires technical rigor and verification, not perfunctory agreement or blind execution
Convert an existing hand-written CLAUDE.md into a typed .spec.ts file for incremental adoption
Write production-grade Rust code using a multi-pass approach. Design types first, then implement, then simplify, then verify with automated lint. Use this skill whenever writing new Rust functions, structs, modules, or features. Triggers on Rust implementation, new Rust code, Rust functions, Rust modules, error handling in Rust, async Rust, or type design in Rust.
Chinese Code Review Guidelines - Provide effective feedback in a way that aligns with domestic team culture while maintaining professionalism and rigor
Run OpenAI Codex CLI as an independent reviewer over the current branch, a specific commit, or uncommitted changes. Builds a focused instruction file from the real diff and returns a compact review summary.