iOS 模擬器測試通常需要可重複的點擊、文字輸入、螢幕截圖與無障礙檢查。這個技能為 Claude、Codex 和 Claude Code 工作流程提供 AXe CLI 指令模式。

Execution plans often pass review while still missing acceptance criteria, deltas, or verification steps, which causes rework during implementation. This skill routes each plan to the correct review lane and checklist, then returns graded blockers so a plan is only marked implementation-ready when it is truly executable.

Vague requests waste engineering time and cause rework. This skill turns them into structured execution plans with clear gates, mandatory steps, and acceptance checks.

Large coding plans often fail when an agent edits too much at once. This skill implements one approved plan step at a time, keeps edits scoped, and reports a clear status when the step is done.

ai-gates documentation can be difficult to apply when project files, plugin files, and hooks have separate roles. This skill explains the workflow boundaries so users can start work, upgrade safely, and avoid duplicate setup.

Substantial feature work can become ambiguous when product intent, technical decisions, and implementation drift apart. This skill guides agents through practical product and technical specifications that stay aligned with shipped work.

Static React interfaces are difficult to generate and maintain from structured data. This skill explains how to render JSON specs, define catalogs, manage state, and connect events in React.

Connecting JSON specifications to Next.js routes, layouts, and server rendering requires several integration steps. This skill guides AI assistants through the documented json-render setup.

JSON-driven interfaces need consistent formatting, calculations, and translated labels. This skill guides directive setup, nested expressions, and custom resolvers for json-render.

Generative interfaces can make rendering errors, state changes, and streamed updates difficult to trace. This skill guides devtools integration and inspection workflows for json-render apps.

AI-generated interfaces need consistent schemas, state bindings, and validation. This skill guides assistants through json-render core APIs for catalogs, prompts, streaming, and specification repair.

Turning UI specifications into maintainable application code can require repetitive traversal and prop serialization. This skill provides reusable utilities for generating structured code and custom exporters from json-render specifications.

Agent development can fail when guidance does not match the installed eve version. This skill directs your assistant to bundled documentation before project changes.

Business conditions can produce unexpected results when field names, missing values, or library behavior differ. This skill builds JsonLogic rules and guides targeted verification.

Large source trees make symbol navigation and caller discovery expensive. This skill guides structural indexing with reusable caches and coverage-aware queries while preserving lexical discovery.

Large changes become difficult to review when unrelated concerns share one branch. This skill guides dependency-based branch planning and non-interactive stacked PR workflows.

Completion claims can hide failed, skipped, or untested requirements. This skill guides requirement-level verification with observed output, commit-bound evidence, and explicit human acceptance.

Implementation reviews lose traceability when findings exist only in conversation. This skill structures task, fix, and branch reviews around contract references and persistent verdict files.

Separate plans can hide missing requirements, incompatible interfaces, and broken dependencies. This skill reviews the complete plan set and coordinates repairs with independent verification.

Incomplete implementation plans leave developers guessing about interfaces, dependencies, and acceptance criteria. This skill structures approved specifications into traceable tasks with explicit tests and execution gates.